Can the United States extend its nuclear umbrella for the protection of its allies against attack, and if so, under what conditions? This is the question that has generated most debate among American and European allied policy makers since the United States began deploying atomic and hydrogen weapons. That the United States would retaliate with its nuclear forces in some large and destructive way in response to an attack on North America with nuclear weapons was not doubted. This case, although socially most destructive, was easy to resolve analytically. There might be some debate, after the Soviet arsenal grew in size and complexity comparable to the American one, about the credibility of American nuclear resolve in the face of less than total attacks on U.S. territory. Still the thrust of U.S., and probably Soviet, strategic analysis assumed that an immediate American nuclear response to a similar attack on U.S. soil was all but assured. The harder cases were those in which the Soviets attacked an American ally to which some semblance of a security guarantee had been offered and to whose protection American power had been committed by treaty or promise. This kind of deterrence was less automatic and more subject to being coerced or tested at the margin.
These cases raise the delicate issue of extended deterrence: the extension of the American nuclear umbrella over Europe or other areas of vital interest assumed to be threatened by Soviet power. Several aspects of this extended deterrence issue will be explored below. The first is its ontology—deciding what exactly it is and can be counted on to provide for American and European defense policy. The second issue is the doctrines and forces on which any strategy of extended nuclear deter-