Here is a little gem I discovered to-day in the files. It is from Kenneth Waltz in his book (1995, The spread of nuclear weapons: A debate) where he engages Sagan (I think it is Sagan) about the benefits of nuclear proliferation:

Second, deterrent balances are inherently stable. This is another reason for new nuclear states to decrease, rather than increase, their military spending. As Secretary Brown saw, within wide limits one state can be insensitive to changes in another state’s forces. French leaders thought this way. France, as president Valery Giscard d’Estaing said, “fixes its security at the level required to maintain, regardless of the way the strategic situation develops in the world, the credibility – in other words, the effectiveness – of its deterrent force.” With deterrent forces securely established, no military requirement presses one side to try to surpass the other. Human error and folly may lead some parties involved in deterrent balances to spend more on armaments than is needed, but other parties need not increase their armaments in response, because such excess does not threaten them. The logic of deterrence eliminates incentives for strategic-arms racing. This should be easier for lesser nuclear states to understand than it was for the United States and the Soviet Union. Because most of them are economically hard-pressed, they will not want to have more than enough. (31)

While I think I may fall more on the Waltz side of the debate, that nuclear proliferation is not as dangerous as we are supposed to believe, I still find some problems with this comment.

The first sentence belies the fundamental assumption Waltz makes, that reality can be measured. To be stable someone needs to know exactly what is happening, and in a security dilemma that observer would need to know what is happening on both sides of the equation. This is an impossibility, because reality is data which needs to be interpreted. Reality does not present warrants. Let us say that France has a minimal deterrent, the stated goal of proliferants according to Waltz. If France were to perceive German arms acquisitions as a capable first strike force then the minimal deterrent is no longer preserved, it is now an insufficient deterrent. A French minimal deterrent also depends upon French interpretations of German willingness to sacrifice. If France thinks German leaders are willing to allow a large population to die, known as willing attrition (which is fair to say the German culture has allowed), then the French minimal deterrence is now not large enough.

Waltz also assumes a security dilemma that is bilateral. Let us assume France (mis)perceives German intentions and reacts accordingly. There are more actors than France and Germany: Russia, England and others might see this move as hostility and as a willingness to first strike or even as a willingness to absorb retaliatory strikes. I think Waltz’ model makes sense in a simple bilateral security dilemma, but that simplicity rarely exists. Maybe it explains the Brazil/Argentina case, but clearly not the India/Pakistan case because China is also involved.

The gap that introduces this problem is actually touched upon by Waltz when he speaks of credibility. Credibility is in the eye of the beholder, what the beholder interprets as an other’s interpretation of the beholder’s actions. Interpretation is too messy and too unscientific to be relied upon.

Waltz also assumes that nations and those security policies are in response merely to exogenous factors. Maybe there is a prestige issue at work and Iranians want to develop nuclear weapons not out of a desire to protect itself or to assert itself, but to (re)establish Persian culture as a major player in the world. US non-proliferation efforts are then seen not as a security action, but rather as a racial action, an attempt by the European Americans to keep down Persians, a recurring story in Persian culture. Waltz’s model neglects these calculations, although this argument’s politics is often in line with Waltz’s politics: allow the proliferation to occur.

Waltz’s defenders have some answers to these arguments, which I will discuss in length later.