Many essays on nuclear strategy and arms control include a discussion of the possible elimination of nuclear weapons. Commentators differ on whether nuclear disarmament would be desirable, but many argue that disarmament is impractical because it could not be verified. Three reasons are often offered for such pessimism. First, nuclear weapons are small and difficult to detect, and one could not be sure that a few weapons had not been hidden away. Second, nuclear weapons are so destructive that a mere handful would confer enormous military and political advantages over non-nuclear adversaries. Finally, nuclear know-how cannot be eliminated, and any nation that had dismantled its nuclear weapons would be capable of quickly assembling a new arsenal from scratch or using civilian nuclear materials. Because of the difficulty of verifying that other states had eliminated all their weapons and providing adequate warning of their rearming, it is argued, states would not agree to disarm in the first place.
While a degree of skepticism is healthy, it is not so obvious that nuclear disarmament could not be adequately verified, particularly in the sort of world in which disarmament would be considered a serious option. The international community, for example, recently concluded that South Africa, which had built a half-dozen nuclear bombs, has disarmed completely and has placed all ofits nuclear materials under international safeguards. Similarly, international inspectors are now confident that they have uncovered all significant nuclear-weapon facilities and activities in Iraq, despite attempts by Iraq to hide such facilities and activities. Moreover, the International Atomic Energy Agency (IAEA) and national intelligence agencies are developing new techniques to find clandestine production facilities and to verify declarations of nuclear material production.
This paper examines the techniques that could be used to verify that nuclear arsenals had been dismantled and to provide timely warning of any attempt to build nuclear weapons. Although no verification regime could provide absolute assurance that former nuclear-weapon states had not hidden a small number of nuclear weapons or enough nuclear material to build a small stockpile, verification could be good enough to reduce remaining uncertainties to a level that might be tolerable in a more transparent and trusting international environment. And although the possibility of rapid break-out will be ever present in modem industrial society, verification could provide the steady reassurance that would be necessary to dissipate residual fears of cheating. Verification will never be so effective that it can substitute for good relations between nations, but it can play an essential role in consolidating the trust that is necessary to support the ongoing process of reducing nuclear arsenals, perhaps all the way down to zero.