Thomas Schelling explains the diplomacy of ultimate survival in Arms and Influence:
As a doctrine, “massive retaliation” (or rather, the threat of it) was in decline almost from its enunciation in 1954. But until 1962 its final dethronement had yet to be attempted. All-out, indiscriminate, “society-destroying” war was still ultimate monarch, even though its prerogative to intervene in small or smallish-to-medium conflicts had been progressively curtailed.
[…]
But in his speech at Ann Arbor, Michigan, in June 1962—a speech reportedly similar to an earlier address in the NATO Council—Secretary McNamara proposed that even in “general war” at the highest level, in a showdown war between the great powers, destruction should not be unconfined. Deterrence should continue, discrimination should be attempted, and “options” should be kept open for terminating the war by something other than sheer exhaustion. “Principal military objectives . . . should be the destruction of the enemy’s military forces, not of his civilian population . . . giving the possible opponent the strongest imaginable incentive to refrain from striking our own cities.”
The ideas that Secretary McNamara expressed in June 1962 have been nicknamed the “counterforce strategy.” They have occasionally been called, as well, the “no-cities strategy.”
[…]
Cities were not merely targets to be destroyed as quickly as possible to weaken the enemy’s war effort, to cause anguish to surviving enemy leaders, or to satisfy a desire for vengeance after all efforts at deterrence had failed. Instead, live cities were to be appreciated as assets, as hostages, as a means of influence over the enemy himself. If enemy cities could be destroyed twelve or forty-eight hours later and if their instant destruction would not make a decisive difference to the enemy’s momentary capabilities, destroying all of them at once would abandon the principal threat by which the enemy might be brought to terms.
[…]
Secretary McNamara incurred resistance on just about all sides. The peace movements accused him of trying to make war acceptable; military extremists accused him of weakening deterrence by making war look soft to the Soviets; the French accused him of finding a doctrine designed for its incompatibility with their own “independent strategic force”; some “realists” considered it impractical; and some analysts argued that the doctrine made sense only to a superior power, yet relied on reciprocity by an inferior power for which it was illogical.
[…]
The idea was not wholly unanticipated in public discussion of strategy; but suggestions by analysts and commentators about limiting even a general war had never reached critical mass. Secretary McNamara’s “new strategy” was one of those rare occurrences, an actual policy innovation or doctrinal change unheralded by widespread public debate. Still, it was not altogether new, having been cogently advanced some 2,400 years earlier by King Archidamus of Sparta, a man, according to Thucydides, with a reputation for both intelligence and moderation.
“And perhaps,” he said,
when they see that our actual strength is keeping pace with the language that we use, they will be more inclined to give way, since their land will still be untouched and, in making up their minds, they will be thinking of advantages which they still possess and which have not yet been destroyed. For you must think of their land as though it was a hostage in your possession, and all the more valuable the better it is looked after. You should spare it up to the last possible moment, and avoid driving them to a state of desperation in which you will find them much harder to deal with.
[…]
The reason for going after the enemy’s military forces is to destroy them before they can destroy our own cities (or our own military forces). The reason for not destroying the cities is to keep them at our mercy. The two notions are not so complementary that one implies the other: they are separate notions to be judged on their separate merits.
[…]
The city-hostage strategy would work best if the enemy had a good idea of what was happening and what was not happening, maintained control over his own forces, could perceive the pattern in our action and its implications for his behavior, and even were in direct communication with us sooner or later. The counterforce campaign would be noisy, likely to disrupt the enemy command structure, and somewhat ambiguous in its target selection as far as the enemy could see. It might also impose haste on the enemy, particularly if he had a diminishing capability to threaten our own cities and were desperate to use it before it was taken away from him.
[…]
Discussions of “counterforce warfare” often imply that the war involves two stages. In the first, both sides abstain from an orgy of destruction and concentrate on disarming each other, the advantage going to the side that has the bigger or better arsenal, the better target location and reconnaissance, the advantage of speed and readiness, and the better luck. At some point this campaign is over, for one side or both; a country runs out of weapons or runs out of military targets against which its weapons are any good, or reaches the point where it costs so many weapons to destroy enemy weapons that the exchange is unpromising.
[…]
This threatened city war is usually implied to be an all-or-none affair, like full-speed collision on the highway, and the driver who has his whole family in the car is expected to yield to the driver who has only part of his family in the car.
[…]
Two adversaries face each other in the knowledge that war is on, each capable of large-scale damage, probably unprecedented damage, possibly damage beyond the ability of either to survive with any political continuity. If each retains more than enough to destroy the other, the counterforce exchange was merely a preliminary, a massive military exercise creating great noise and confusion (and undoubtedly great civilian damage too), but constituting an overture to the serious war that is about to begin.
[…]
We know little about this kind of violence on a grand scale. On a small scale it occurs between the Greeks and the Turks on Cyprus and it occurred between the settlers and the Indians in the Far West. It occurs in gang warfare, sometimes in racial violence and civil wars. Terror is an outstanding mode of conflict in localized primitive wars; and unilateral violence has been used to subdue satellite countries, occupied countries, or dissident groups inside a dictatorship. But bilateral violence, as a mode of warfare between two major countries, especially nuclear-armed countries, is beyond any experience from which we can draw easy lessons.
There are two respects in which a war of pure violence would differ from the violence in Algeria or Cyprus. One is that insurgency warfare typically involves two actively opposed sides—the authorities and the insurgents—and a third group, a large population subject to coercion and cajolery. Vietnam in the early 1960s was less like a war between two avowed opponents than like gang warfare with two competing gangs selling “protection” to the population.
There is a second difference. It involves the technology of violence. Most of the violence we are familiar with, whether insurgency in backward areas or the blockade and strategic bombing of World Wars I and II, were tests of endurance over time in the face of violence inflicted over time. There was a limit on how rapidly the violence could be exercised. The dispenser of violence did not have a reservoir of pain and damage that he could unload as he chose, but had some maximum rate of delivery; and the question was who could stand it longest, or who could display that he would ultimately win the contest and so persuade his enemy to yield. Nuclear violence would be more in the nature of a once-for-all capability, to be delivered fast or slowly at the discretion of the contestants. Competitive starvation works slowly; and blockade works through slow strangulation. Nuclear violence would involve deliberate withholding and apportionment over time; each would have a stockpile subject to rapid delivery, the total delivery of which would simply use up the reserve (or the useful targets).
[…]
If I waylay your children after school, and you kidnap mine, and each of us intends to use his hostages to guarantee the safety of his own children and possibly to settle some other disputes as well, there is no straightforward analysis that tells us what form the bargaining takes, what children in our respective possessions get hurt, who expects the other to yield or who expects the other to expect oneself to yield—and how it all comes out.
[…]
This is a strange and repellent war to contemplate. The alternative once-for-all massive retaliation in which the enemy society is wiped out as nearly as possible in a single salvo is less “unthinkable” because it does not demand any thinking.
[…]
And it may seem less cruel because it does not have a cruel purpose—it is merely purposeless—compared with deliberate, measured violence that carries the threat of more. It does not require calculating how to be frightful, how to terrorize an adversary, how to behave in a fearsome way and how to persuade somebody that we are more callous or less civilized than he and can stand the violence and degradation longer than he can.
[…]
In earlier times, one could plan the opening moves of war in detail and hope to improvise plans for its closure; for thermonuclear war, any preparations for closure would have to be made before the war starts.
Even the enemy’s unconditional surrender might be unavailable unless one had given thought in advance to how to accept and to police surrender. And a militarily defeated enemy, desperate to surrender, might be unable to communicate its offer, to prove itself serious, to accept conditions and to prove compliance on the urgent time schedule of supersonic warfare unless it had given thought before the war started to how it might end.
[…]
With today’s weapons it is hard to see that there could be an issue about which both sides would genuinely prefer to fight a major war rather than to accommodate. But it is not so hard to imagine a war that results from a crisis’ getting out of hand.
[…]
Bombers vividly illustrate the dynamic character of war—the difficulty of finding resting places or stopping points, the impossibility of freezing everything. Bombers cannot just “stop.” They have to move to stay aloft; to move they have to burn fuel, and while they are moving crews become fatigued, enemy defenses may locate and identify them, and coordination of the bombers with each other will deteriorate. If the planes return to base they have to recycle, with refueling and other delays if the truce was a false one and the war is still on. Finally, at base they may be vulnerable. The aircraft would have been launched quickly for their own safety, and once in the air were immune to missile attack; when they return to base they may be newly vulnerable—if the bases still exist. If they have to seek alternative bases their performance would be further degraded, and so would the threat they pose to the enemy in bringing about, or keeping, a precarious armistice.
[…]
There are some good reasons for supposing that, if the war could be stopped, it might be a simple cease-fire that would do it. Considering the difficulty of communication and the urgency of reaching a truce, simple arrangements would have the strongest appeal and might be the only ones that could be negotiated on the demanding time schedule of a war in progress. And a crude cease-fire might be the only stoppage that could be arrived at by tacit negotiation, by the mere extension of a pause.
[…]
(The side first motivated to announce its terms could be either the stronger or the weaker, the one most hurt or the one least hurt, the one with the most yet to lose or the one with the least yet to lose, the one that started the war or the one that did not—and it might not be clear who started it, who had been hurt worse, or who eventually had the most yet to lose.)
[…]
If among the terms of the understanding were that one side should disarm itself, partially or completely, of its remaining strategic weapons, ways would have to be found to make it feasible and susceptible of inspection. If it had to be done in a hurry, as it might have to be, enemy aircraft could be required to land at specified airfields, even missiles could be fired at a point where their impact could be monitored (preferably with their warheads disarmed or removed), and submarines might surface to be escorted or disabled. Certainly of all the ways to dispose of remaining enemy weapons, self-inflicted destruction is one of the best; and techniques to monitor it, facilitate it, or even to participate in it with demolition charges would be better than continuing the war and firing away scarce weapons at a range of several thousand miles.
“Uncontested reconnaissance” would be an important part of the process. Submitting to surveillance, restricted or unrestricted, might be an absolute condition of any armistice. In the terminal stage of the war, it is not just “armed reconnaissance” that could be useful but “unarmed reconnaissance,” uncontested reconnaissance by aircraft or other vehicles admitted by sufferance.
[…]
If one side is submitting to a very asymmetrical disarmament arrangement, it may have to prove how strong it is for bargaining purposes and then prove how weak it is in meeting the disarmament demands of its opponent. For purposes of bluff it would be valuable to have an opponent think one had hidden weapons in reserve; for abiding by a truce arrangement it may be frustrating and dangerous to be unable to deny convincingly the possession of weapons that one actually does not have.
A critical choice in the process of bringing a war to a successful close—or to the least disastrous close—is whether to destroy or to preserve the opposing government and its principal channels of command and communication. If we manage to destroy the opposing government’s control over its own armed forces, we may reduce their military effectiveness. At the same time, if we destroy the enemy government’s authority over its armed forces, we may preclude anyone’s ability to stop the war, to surrender, to negotiate an armistice, or to dismantle the enemy’s weapons. This is a genuine dilemma: without technical knowledge of the enemy’s command and control system, the enemy’s war plan and target doctrine, the vulnerabilities of enemy communications and the procedures for implementing military action, we cannot reach a conclusion here. All we can do is to recognize that there is no obvious answer. Victorious governments have usually wanted to deal with an authority on the other side that could negotiate, enter into commitments, control and withdraw its own forces, guarantee the immunity of ambassadors or surveillance teams, give authoritative accounts of the forces remaining, collaborate in any authentication procedures required to verify the facts, and institute some kind of order in its own country. There is strong historical basis for presuming that we should badly want to be sure that an organized enemy government existed that had the power to demand its armed forces cease, withdraw, submit, mark time, or perform services for us. This has to be weighed against the advantage of disorganizing the initial enemy attacks by destroying the enemy command structure.
[…]
If we were certain that he would fire all of his weapons as quickly as he could, and fire them to maximize civilian damage on our side, the argument for going after his weapons quickly and unstintingly would be conclusive. If alternatively we were certain that he preferred to pause and negotiate, but nevertheless would fire his weapons rather than see them destroyed on the ground, our all-out attack on them would simply pull the trigger; the argument against it would then be conclusive.
[…]
To think of war as a bargaining process is uncongenial to some of us. Bargaining with violence smacks of extortion, vicious politics, callous diplomacy, and everything indecent, illegal, or uncivilized. It is bad enough to kill and to maim, but to do it for gain and not for some transcendent purpose seems even worse. Bargaining also smacks of appeasement, of politics and diplomacy, of accommodation or collaboration with the enemy, of selling out and compromising, of everything weak and irresolute. But to fight a purely destructive war is neither clean nor heroic; it is just purposeless. No one who hates war can eliminate its ugliness by shutting his eyes to the need for responsible direction; coercion is the business of war. And someone who hates mixing politics with war usually wants to glorify an action by ignoring or disguising its purpose. Both points of view deserve sympathy, and in some wars they could be indulged; neither should determine the conduct of a thermonuclear war.
What is the bargaining about? First there is bargaining about the conduct of the war itself. In more narrowly limited wars—the Korean War, or the war in Vietnam, or a hypothetical war confined to Europe or the Middle East—the bargaining about the way the war is to be fought is conspicuous and continual: what weapons are used, what nationalities are involved, what targets are sanctuaries and what are legitimate, what forms participation can take without being counted as “combat,” what codes of reprisal or hot pursuit and what treatment of prisoners are to be recognized. The same should be true in the largest war: the treatment of population centers, the deliberate creation or avoidance of fallout, the inclusion or exclusion of particular countries as combatants and targets, the destruction or preservation of each other’s government or command centers, demonstrations of strength and resolve, and the treatment of the communications facilities on which explicit bargaining depends, should be within the cognizance of those who command the operations. Part of this bargaining might be explicit, in verbal messages and replies; much of it would be tacit, in the patterns of behavior and reactions to enemy behavior. The tacit bargaining would involve targets conspicuously hit and conspicuously avoided, the character and timing of specific reprisals, demonstrations of strength and resolve and of the accuracy of target intelligence, and anything else that conveys intent to the enemy or structures his expectations about the kind of war it is going to be.
Second, there would be bargaining about the cease-fire, truce, armistice, surrender, disarmament, or whatever it is that brings the war to a close—about the way to halt the war and the military requirements for stopping it. The terms could involve weapons—their number, readiness, location, preservation, or destruction—and the disposition of weapons and actions beyond recall or out of control or unaccounted for, or whose status was in dispute between the two sides. It would involve surveillance and inspection, either to monitor compliance with the armistice or just to establish the facts, to demonstrate strength or weakness, to assign fault or innocence in case of untoward events, and to keep track of third parties’ military forces. It could involve understandings about the reassembling or reconstituting of military forces, refueling, readying of missiles on launching pads, repair and maintenance, and all the other steps that would prepare a country either to meet a renewed attack or to launch one. It could involve argument or bargaining about the degree of destruction to people and property on both sides, the equity or justice of what had been done and the need to inflict punishment or to exact submissiveness. It could involve the dismantling or preservation of warning systems, military communications, or air defenses. And it very likely would involve the status of sheltered or unsheltered population in view of their significance as “hostages” against resumption of warfare.
A third subject of bargaining could be the regime within the enemy country itself. At a minimum there might have to be a decision about whom to recognize as authority in the enemy country or with whom one would willingly deal. There might be a choice between negotiating with military or civilian authorities; and if the war is as disruptive as can easily be imagined, there may be a problem of “succession” to resolve. There could even be competing regimes in the enemy country—alternative commanders to recognize as the inheritors of control, or alternative political leaders whose acquisition of control depended on whether they could monopolize communications or get themselves recognized as authoritative negotiators. To some extent, either side can determine the regime on the other side by the process of recognition and negotiation itself. This would especially be the case in the decision to negotiate about allied countries—China, or France and Germany—or alternatively to refuse to deal with the primary enemy about allied and satellite affairs and to insist upon dealing separately with the governments of those countries.
A fourth subject for bargaining would be the disposition of any theater in which local or regional war was taking place. This could involve the evacuation or occupation of territory, local surrender of forces, coordinated withdrawals, treatment of the population, use of troops to police the areas, prisoner exchanges, return or transfer of authority to local governments, inspection and surveillance, introduction of occupation authorities, or anything else pertinent to the local termination of warfare.
[…]
Fifth would be the longer term disarmament and inspection arrangements.
[…]
For that reason the armistice might, as in the days of Julius Caesar, involve the surrender of hostages as a pledge for future compliance.
[…]
A sixth subject for negotiation might be the political status of various countries or territories—dissolution of alliances or blocs, dismemberment of countries, and all the other things that wars are usually “about,” possibly including economic arrangements and particularly reparations and prohibitions.
[…]
We are dealing with a process that is inherently frantic, noisy, and disruptive, in an environment of acute uncertainty, conducted by human beings who have never experienced such a crisis before and on an extraordinarily demanding time schedule. We have to suppose that the negotiation would be truncated, incomplete, improvised, and disorderly, with threats, offers, and demands issued disjointedly and inconsistently, subject to misunderstanding about facts as well as intent, and with uncertainty about who has the authority to negotiate and to command.
[…]
In ordinary peacetime the Soviet leaders have tended to disdain the idea of restraint in warfare. Why not? It permits them to ridicule American strategy, to pose the deterrent threat of massive retaliation, and still perhaps to change their minds if they ever have to take war seriously. On the brink of war they would. It may be just before the outbreak that an intense dialogue would occur, shaping expectations about bringing the war to a close, avoiding a contest in city destruction, and keeping communications open.