Hot on the heels of the development of nuclear weapons, strategies for dealing with them -- bad pun alert! -- mushroomed. Much of it emanated from the RAND Corporation, home to, as Fred Kaplan explained in The Wizards of Armageddon (Touchstone, 1983) "a small and exceptionally inbred collection of men -- mostly economists and mathematicians, a few political scientists -- who devoted nearly every moment of their workaday thoughts to thinking about the bomb: how to prevent nuclear war, how to fight nuclear war if it cannot be deterred."
Specific subjects included fun stuff like first and second strikes, the always popular mutual assured destruction (MAD), launch on warning, and, finally, targeting cities versus targeting "counterforce" (the enemy's nuclear weapons).
Daunting as that sounds, some of the concepts are deceptively simple. Or, to put it another way, they started out simple, but were worried to death in think tanks and other institutions. For example, the term deterrence, when used in international relations, just means using the threat of an attack to compel a foe to either act or refrain from acting in accordance with the deterring state's wishes. Yet, applied to nuclear weapons, deterrence has spawned countless books, papers, and conferences.
An arms race, of course, is the principal component of deterrence: each side tries to stockpile weapons more advanced than those of its designated enemy to keep it from attacking. In the field of nuclear weapons, the arms race manifested itself as, for example, development of first, the atomic bomb, then, the exponentially more powerful hydrogen bomb; first, bombers, then intercontinental missiles.
The second element of an arms race besides maintaining a development edge over one's enemy is building more weapons. Human nature, right? When it comes to nuclear weapons, one assumes that the perceived need to go forth and multiply is determined by how many weapons the enemy has. In fact, it's more dependent on the number of your enemy's targets. In the January/February issue of the Bulletin of the Atomic Scientists, Ivan Oelrich explained how this works (behind a pay wall).
Nuclear weapons were first aimed at cities, the centers of population and industrial production. This targeting strategy was, in part, a natural extrapolation of the mass city bombing tactics of World War II. But cities also became nuclear targets by default; with the very inaccurate early missiles, cities were the only targets big enough to hit. As accuracy, weapon numbers, and intelligence on enemy-weapon location increased, it was irresistible for both sides to target the enemy’s nuclear weapons. … The targeting of weapons inevitably led to an arms race. If cities were the only targets, then neither side needed more weapons than the other side had cities to shoot at. But once nuclear weapons became targets, each side had to have as many weapons as the other side for counterforce attacks, plus more to shoot at “value” targets like cities. When each side needed just a few more than the other, an arms race without end was on. [Emphasis added.]
One's first impression is that attacking weapons is significantly less barbaric than attacking population centers. Alas, in practice, it doesn't work that way. In its counterintuitiveness, switching a nation's nuclear weapons policy away from counterforce parallels missile defense. In the first, you're restraining yourself from taking away your enemy's weapons. In the second, you surrender the ability to shoot down attacking warheads to keep from inciting your enemy to make more to both overwhelm your missile defense and mount a second attack if the first is thwarted.
That's part of what makes nuclear races so lethal: the field is fraught with seeming Sophie's Choices like that.