A specter in the US military circles during the late phase of the cold war. In the early 80s, with the introduction of the SS-24 Scalpel and the SS-25 Sickle, the USSR was about to acquire a new generation of inter-continental ballistic missiles with sufficient accuracy to target individual missile silos. At the same time, the use of MIRV meant that each silo contained several nuclear bombs and each missile could destroy several silos. The net result being that the USSR could in theory destroy the entire land-based US missile force using only a fraction of their own.
At least in the eyes of American decision-makers, this created an interesting dilemma. What if the Soviets launched a first strike that spared the civilian population, but destroyed the US ICBM-force? The other two legs of the nuclear triad would remain, i.e. the bombers and the missile submarines, but bombers are ineffective without missiles to destroy the enemy air-defenses, and the SLBMs of the time were only accurate enough to hit cities, not military targets. Thus the only two options would be to launch a retaliatory strike against civilian targets, which would be suicidal since it would force Soviet retaliation in turn, or to acknowledge defeat.
The disabling first strike was used as motivation for several expensive projects, notably the MX Missile project that became the LGM-118A Peacekeeper, and the very accurate Trident II D-5 SLBM. But the assumptions behind the concept have been challenged. First, it seems doubtful that ICBM force could really be wiped out completely - at least with such certainty that the USSR would dare to base any calculations on it. A conservative estimate might be that 10% would remain, a quite significant force.
More importantly, the clean separation between military and civilian targets that would be required to destroy the former without hurting the latter simply does not exist. An independent study1 made for the US Congress about the effects of nuclear war estimated between 1 and 20 million short-term deaths from a "limited" attack on ICBM fields only. The variation depends on differing assumptions about weapons used and available fallout protection, but "deaths below the 8 to 10 million level requires quite optimistic assumptions". This should be compared to between 20 and 165 million in an all-out war. It seems improbable that the US leaders would let such a blow pass without retaliating, and next to impossible that the USSR would calculate with them doing so. Furthermore, such an attack would presumably also severely damage the Command & Control structures needed to distinguish between a "limited" and an "all-out" scenario and decide on the appropriate response, introducing even more uncertainty.
Nigel Calder, in his book Nuclear Nightmares2, has described the first-strike hysteria as a cultural misunderstanding due to differences of military strategy in the USA and the USSR. The USSR had in fact been aiming ICBMs at ICBMs all along since the 60s, as evidenced by their producing a lot more megaton-warheads than there are American cities worthy of such a charge - if you cannot decrease the radius of error of your missile you have to increase the radius of destruction. At no time did they have sufficiently many to completely destroy the US force, but in the event of an imminent attack they would be able to destroy a fraction of it, and each warhead destroyed on American soil would be one less to detonate over the Soviet Union.
This difference was representative of the difference in attitude to nuclear weapons in general. The USA tended to think of the ICBM as an ultimate weapon against which there was no defense, and instead targeted enemy cities to deter against attack. In the USSR, the Bomb was regarded as a powerful weapon among others, and plans were made to maximize its effect and to defend against its use by the enemy. Thus the anti-missile force was gradually ramped up. From this point of view, the introduction of more accurate missiles was simply a economic measure that enabled the use of a greater number of charges with lower yield. But at the same time it made it obvious that the target was ICBM silos and not cities. America saw, and panicked.
The existence of a counterforce force is of course troubling in itself, because it inherently acts destabilizing on the mutual deterrence. In a crisis it can lead to a situation where it appears better to launch a first strike than to wait, because it would lessen the impact of the opponent's strike or because it lets you use missiles that seem about to be destroyed. But the aim of the Soviet counterforce capability was defensive, not offensive, just like the counter-force capability that the USA developed in response. And so the arms race goes through another cycle.
references:
- The Effects of Nuclear War. Office of Technology Assessment, 1979.
Available online:
http://www.wws.princeton.edu/~ota/disk3/1979/7906_n.html
- Nigel Calder, Nuclear nightmares : an investigation into possible
wars. Harmondsworth : Penguin, 1981