In the my blog post on Australia’s decision to buy a fleet of Joint Strike Fighters, I mention the security dilemma. The following is taken from a draft of my second, as yet not published, book on weapons research – with deleted bits marked with ellipses and interpolated bits in [..]. The context is a discussion about weapons research and deterrence and takes place near the end of the book. I claim by that stage to have shown that there is no such thing as a weapon that is ‘inherently defensive’ in the sense that it can only be used to resist aggression and cannot play any possible role in an aggressive war – I will say more about this on the site later on. My aim in the present discussion is to indicate how dangerous it is to acquire weapons for deterrence.

Deterrence can induce an ‘action-reaction’ cycle… when it involves weapons research: B engages in weapons research to produce the means to deter A from d, A never intended to do d, so interprets this as an aggressive move by B and so A also engages in weapons research, and so on. Robert Jervis … wrote about the security dilemma in the Cold War in a similar vein. [State] B feels insecure (possibly also believing it must deter A) and so acquires a new military capability; becoming aware of this, A now feels insecure, and it too acquires new capabilities, and so on [this is the security dilemma]. Jervis, as we saw, suggested that developing defensive systems would not induce this cycle of arms acquisition, because defence is not threatening - …The security dilemma, and the attendant action-reaction arms race, most certainly can be generated by the belief that deterrence is needed. And this is another reason why deterrence is dangerous. We can also remark here that weapons research takes time, especially for complicated systems like nuclear weapons. So to undertake weapons research for deterrence, the assumption must be that deterrence will be necessary in several years hence, when the work will be completed. It is clear that this too will tend to fuel an ongoing cycle and commitment to both deterrence and the weapons research believed necessary to ensure the status quo. Ongoing programmes must continue to be initiated, or it may turn out in several years time that the means for deterrence are unavailable.

Now we come to what I take to be the conclusive reason why deterrence theory cannot justify weapons research. Deterrence is a relationship that takes place for a finite time; states do not deter one another for ever, as the end of the Cold War has demonstrated. But the products of weapons research are not like that: they live on, perhaps for all time, as do the designs for all the nuclear systems developed during the Cold War. The need for deterrence has passed, but the means for deterrence remains, posing a danger into the distant future. The pressing problem after 1965 was nuclear deterrence, how to ensure it and what would happen if it failed. The means for deterrence by that time assured the destruction of both parties, so the price of deterrence failing was incomparably bad. When there is no longer a need for deterrence, the means for deterrence still remain and the possibility remains open that some time they will be used. It follows that the general criterion for justifying violations of moral rules cannot be satisfied. It can never be shown that weapons research done in order to ensure deterrence, in a given historical context, will not cause harms at some time in the future, long after the need for deterrence has passed.