After more than 50 years and over $100 billion spent on research and development, missile defense is the longest running and most expensive military research program in U.S. history. This dissertation examines why policymakers disagree on whether a missile defense technology can be built and why their disagreements proceed unabated in spite of changes in the strategic environment and the state of relevant technology.
Through archival research and interviews with members of the policymaking community, this dissertation evaluates these questions in the context of X-ray laser technology considered from 1983 until 1988 and Hit-to-Kill technology during the period 1996 to 2007. Empirical analysis reveals that policymakers' intractable differences on whether an NMD system can be built emerge and persist due to the presence of irreducible uncertainty. No one can know exactly what will be required of a missile defense system or what can be done. The required capability hinges on the unknowable future capability and intent of a potential adversary and the U.S. capability rests on the significance of limited existing data for future technical developments, forced to square this circle, and proceed from irreducible uncertainty to finite technical judgments, scientists, government officials and policy experts rely on the content of their beliefs about strategic prudence in a nuclear environment to define what technology can and must be able to do.
Those who believe that an NMD system will enhance U.S. security contend that it can be built and policymakers who believe that an NMD system will hurt U.S. security find that the myriad feats cannot be accomplished. NMD Supporters' "War Fighting'' belief structure, which emphasizes the continued relevance of the security dilemma, leads them to conclude that a missile defense system can be built by minimizing the importance of what an adversary may do and by maximizing the significance of any progress exhibited in existing data. However, Opponents' "Arms Control" belief structure points to the overriding significance of mutual vulnerability for state security and leads them to determine that an NMD system cannot be built by maximizing the importance of an adversary's future capability and intent as well as the significance of failures and limits in existing data.
