This paper will analyze literature related to theories of nuclear deterrence and will seek to examine (1) historical application, (2) scientific and logical validity, and (3) the relevance of deterrence theories for international relations. I hypothesize that while deterrence theory has been widely used and relied upon, particularly in the context of relationships between modern nuclear powers, there are serious scientific and logical concerns as to the true validity of deterrence theories. These include problems with (1) circular logic; (2) bias; (3) difficulty in isolating the independent variables; (4) validation and falsification; and (5) misapplication or misinterpretation in scenarios where disparity exists between levels of risk aversion among actors.

HISTORICAL APPLICATION OF DETERRANCE THEORY

In his chapter entitled “The Anatomy of Deterrence,” Bernard Brodie reflects on the ‘romantic’ notion that peace can be achieved through a negotiated peace as a result of military action on the part of strong and decisive leaders. (Brodie, 266). However, he makes the argument that in an environment where nuclear weapons of mass destruction exist, such military actions could (and most likely would) lead to a level of military conflict that would prove cataclysmic to human civilization as we know it. Therefore, a different understanding and explanation of international relations is called for.

In his writings, Brodie advocates the strategies of deterrence and limited war which would, in theory, lead to levels of power parity between nuclear parties on the international stage. Therefore, an emphasis is placed on creating a scenario whereby any actor with the technology, ability, and will to launch a first-strike offensive attack on another nuclear nation would be met with the sure and certain fact that the other nation would, without restraint, launch a full-scale second-strike attack, leading to the destruction of both nations and, most likely, a large portion of human civilization as we know it. The premise of deterrence theory, then, is similar to that of a grand-scale game of “chicken” whereby each party desires to convince the other that they will not back down from a straight-on attack and are willing to destroy the other party and themselves in the process.

Building upon the realist and neo-realist assumptions of unitary state actors, rationality and/or bounded rationality, and the idea that balances in power lead to a lower probability of conflict, this theory calls for the one for one buildup of nuclear arms with any nation capable of launching a first-strike attack, in order to openly demonstrate a nation’s willingness and capability to launch a second-strike retaliatory action against any first-strike attack.

While this theory would appear, on its face, to be a natural extension of neo-realism and its theories, it would also appear to differ in some of its logic. For instance, in this case, a nation is hedging its future on the belief that it can convince the other nuclear power(s) that it is willing to “guarantee not only vast losses but also utter defeat” (Brodie, 276) at its own expense, if necessary. So, unlike traditional neo-realism, the idea of deterrence doesn’t simply rely on the underlying assumption that nations seek to maximize power, but also assumes that nations seek to defend themselves through a “tit for tat” approach to the buildup of power and capability that they actually hope never to use.

ASSESSING THE VALIDITY OF THE DETERRENCE THEORY

A major criticism of traditional deterrence theory must be that it seeks to openly convince another nation that it will do something it clearly would not rationally choose to do, that is destroy itself in order to counter an offensive attack from the other side. The logic in this theory becomes almost circular in its approach. A nation is assumed to build-up power in order to defend itself from another nation that is building up power in order to defend itself from the first nation’s buildup of power … and on it will go. So, it would appear through the logic of deterrence that there is actually an incentive to buildup power on both sides in a way that actually leads to their possible demise. This is best explained by way of the “security dilemma” which would say that as each side seeks to build-up its nuclear capabilities in response to the other, both sides actually increase the probability that a conflict will occur. When viewing the buildup of nuclear capabilities, this should be seen from the outside observer as an “irrational” act, yet each side, seeking to maximize its power and military capability, sees the buildup as a rational act.

However, one must question not only the logic of this theory but also the entire premise that a theory such as this can attempt to explain a “non-event.” In other words, this theory really depends upon events not occurring in order to demonstrate its reliability. If something must not occur in order for a theory to be reliable, how can that theory be tested or falsified? As Jervis (P. 293) observes, “explaining basic changes in the relationship between states is beyond the boundaries of the theory … thus diminishing its validity.” In other words, we only can see something that didn’t happen empirically, but we really don’t know what might have happened if the scenario had been different; and, as we discussed in last week’s lecture, there most likely would be no one around to report the results of an actual event or falsified observation which would lead us to empirically say that the theory is invalid. Therefore, we must rely on assumption and generalization to say that deterrence is the true reason that a nuclear confrontation between superpowers never occurred.

In addition to examining the circular logic of deterrence theory, one must also consider the presence of intervening variables, which might spuriously impact the outcome of a diplomatic or international event. For instance, in the obvious example of the Cuban Missile Crisis, was it deterrence that led the Soviets to back down from their decision to implement strategic missiles in Cuba or was it a combination of diplomatic channels, internal domestic differences between decision makers, military strategic decisions, or just pure “luck?” One could argue that any one of these variables may have been the deciding factor in the end result of the conflict. In other words, we don’t know exactly what effect deterrence had on the outcome because it would be difficult, if not impossible, to control for it and/or falsify the outcome.

Another observation, made by Jervis (P. 297) is that deterrence theory is biased in favor of status-quo powers. In other words, the status-quo powers may have the desire and/or ability to coerce others to behave in a manner that favors the status quo power. Jervis and others argue that the game of chicken is not an appropriate explanation of nuclear deterrence in that it assumes that “both cars are in motion at the start.” However, if we assume that one car got a head start on the other, the other car is more apt to make a strategic mistake that can lead to devastating results. If the two powers are not equally risk averse, and if the power that is not satisfied with the status-quo is the party that is less risk averse, a higher probability of nuclear exchange would be the result. This scenario would critically harm the validity of deterrence theory. And, as was stated earlier, if the probability exists that deterrence theory is, in any way, possibly invalid, it would be illogical and possibly disastrous to depend upon it when making decisions in the international arena.

The discussions contained within this paper lead to an even more important criticism. What exactly is rational behavior? It would appear, as in the criticism of circular logic, that the rational decision made by a unitary actor desiring to maximize power would most likely be seen from the outside as an irrational decision, as it has a high probability of harming the very nation that is attempting to preserve itself. Therefore, a serious criticism of traditional deterrence theory, as outlined by Jerves (P. 299), is that a well stated and objectively measured operational definition of rational behavior does not exist within the literature.

Is it rational to make a decision that would lead to the demise of one’s own civilization? Or is it more rational to assume that if a first strike is launched against a nation, and the nation’s destruction is inevitable, then actions that will lead to the destruction of the other civilization are the logical choice? In this case, rationality is relative to perceived risk and to the risk-aversive behavior of both parties. However, if either party’s level of risk aversion is below some unpredictable point, the assumption of rationality becomes invalid, as would the entire theory.

In defense of Jervis’ work in addressing long-term stability under the deterrence model, Kugler and Zagare (1990) attempt to extend the traditional model by combining it with the model of power transitions, and make the observation that parity between nuclear parties is not stable, but rather is a critical point where conflict and/or nuclear exchange can occur. Again, since no specific exchange of nuclear weapons has taken place under these circumstances to date, this theory is far from verified and is definitely not in a position that can be defended through experimentation or falsification. Therefore, one must criticize their work as broad-based assumptions made by applying the work of other authors within the field. However, it would seem logical, and somewhat practical to believe that there is some glimmer of truth in the arguments made within Kuglar and Zagare’s paper. Therefore, while their work is not at a position that could be called scientifically progressive under the standards established by Lakatos, it can also be said that their theories are not degenerative to the previous models as they do not alter the hard-core assumptions of unitary actors and rational decision making. Therefore, my criticism of this work is far less negative than my criticism of the traditional model of deterrence outlined by Brodie.

An additional observation is that Kuglar and Zagare’s work adds to the second wave models described by Jerves in that it examines the phenomenon of risk aversive actors and uses game theory to further determine the possible probabilistic outcomes of confrontational scenarios. But, as with the other explanations, we must depend upon the probability of predicting a non-event (long-term stability measured by the absence of a nuclear exchange) versus a catastrophic event (nuclear exchange). Either way the model is un-testable and unverifiable. It is more of an explanation than a true scientific model or theory. Thus, we are relying on assumptions and generalizations.

In the works of Signorino and Tarar (2006) and Zagre (1990), we again see the application of game theory as a way of empirically explaining and/or predicting the possible outcomes of nuclear buildups by examining risk and uncertainty on the part of “defenders” and “attackers.” Again, we can see that if the attacker is the party that is less risk-averse, the probability of war should, theoretically, increase, thus explaining that there is really no such thing as “long-term stability” when evaluating nuclear deterrence. While the application of probit analysis to the model does add an elegant mathematical presentation, it is still not a progressive model that can be verified experimentally or falsified scientifically, as the results of such experimentation would more than likely prove disastrous to the entire world. Therefore, the criticism of these models remains that they cannot be empirically analyzed in a way that would satisfy the requirements of even the loosest constraints placed by the Lakatosian system of evaluation.

While Signorino and Tarar (p. 592) claim that their model explains 96% of the attackers actions and over 93% of outcomes using an analysis of the POLITY III data set from 1800-1994, the problem still remains that they are trying to empirically predict a nuclear exchange using non-nuclear conflicts and therefore are not addressing some very important elements. For instance, they are not really analyzing the willingness of either party to truly use nuclear weapons in a first or second-strike position. And, given that the United States is the only nation to ever definitively make the decision to use a nuclear weapon in a conflict, we must also critically examine their application of democratic peace theory to the analysis of deterrence theory. Thus, even the most contemporary of the papers reviewed for this analysis do not truly bring us to a model of deterrence that can effectively be used to explain or predict the actions of nuclear parties. One additional criticism which must be considered is that none of the literature assigned for this class addresses non-state actors. Given our post 9-11 “war on terror,” it is vital that we examine our current geopolitical environment in future research.

CONCLUSIONS

A review of literature assigned for our study of deterrence theory and a critical analysis of their findings leads me to believe that, while deterrence theory has a logical basis for explanation, it lacks a level of scientific validity which would allow us to depend entirely on its hypotheses and hard core assumptions. Therefore, reliance upon this theory as the sole explanation for the behavior of nuclear states and for the buildup of nuclear weapons in response to the actions of other states is not scientifically supported. Other methods such as negotiation, arms limitations, treaties, and efforts to promote sustainable economic growth among nations are also required for any nation that wishes to contribute to long-term stability in the modern nuclear age.