About one month ago I jumped on the Substack bandwagon wanting my own space to write and rant about all things deterrence. Until today one thing or another popped up and the Substack remained low on the priority list. However, scrolling through my Twitter feed yesterday afternoon I came across a tweet from one of my favorite academics, Sergey Radchenko, in which he has some very nasty things to say about self-deterrence:
Having spent a decent amount of time exploring this regrettably understudied concept, and recently presented a paper on it at the ISA Conference in Montreal, I took some offense at the idea that “There’s no such thing as self-deterrence”. Immediately upon reading Sergey’s tweet, I decided to reorder my list of priorities and prepare a response.
In fairness to Sergey, the term 'self-deterrence’ has been employed quite often by various IR and strategic studies scholars since at least the late 1960s. Donald Brennan, Colin Gray, Robert Jervis, Richard Ned Lebow, T.V. Paul and Lawrence Freedman are among the scholars who have used the term in their writings on deterrence. However, none of these superb scholars have dedicated more than a few pages to explaining what the concept refers to or provided illustrations, and for this oversight they should be ashamed of themselves!
Frankly, whether the term 'self-deterrence’ is employed, or some similar term, such as ‘internalized deterrence’, is irrelevant. What matters is that there is a conceptual gap in deterrence research and it still needs to be properly filled.
You might be saying to yourselves, ‘A gap in deterrence research? Pull the other one!’ Seriously, I really believe there is an important gap.
From the origins of the field to the present day, the overwhelming preference of policymakers, defense officials and scholars is to discuss deterrence in terms of deterring others (how does A deter B). This seems natural. We want to identify the most effective ways and means of deterring our adversaries. Why bother then thinking in terms of how we might deter ourselves? Largely due to this sort of mindset, hardly any attention has been devoted to the problem of self-deterrence (in what ways does B deter B), despite the fact there is presumably some merit in identifying how our adversaries deter themselves. Indeed, the problem of how political and military strategies may wittingly or unwittingly account for the self-deterrence of adversaries (how is the formulation of A’s strategy shaped by, or can take advantage of, B deterring B), has also received far too little attention.
Where I take issue with Sergey and other doubters is that they need to zoom out before they zoom in. In other words, let’s park self-deterrence for a moment and consider what is meant by a ‘deterrence effect’, or to put it slightly differently, the effect of being deterred. Where does this effect come from? Traditionally, this effect is understood to come about when A threatens B and B is deterred from taking a course of action A has communicated to them not to take. Naturally this raises a separate set of questions about how one identifies this ‘deterrence effect’ otherwise known as a ‘deterrence success’, in actual cases. I will address this longstanding problem in a separate post in due course. For now let us stay at an abstract level. Is there something obvious that is missing from this formulation? We’ve got actors A and B. We’ve also got A’s threat of consequences, A’s identification of the action it doesn’t want B to take, and then B’s reaction to it.
Let’s focus on the threat a moment. For deterrence theorists, the threat is key. If A’s threat is insufficient it won’t deter B. Therefore it is essential for A to have the sufficient means to threaten, the credibility to be taken seriously, and the ability to effectively communicate its threat to B. If all goes well, if all these boxes are checked, then deterrence is likely to be successful, or at least have a reasonable chance of success. Right? Not so fast.
At least four things are missing. First, it is necessary to differentiate between A’s intended threat and B’s reception and interpretation of A’s threat. The more ambiguous the threat, for instance, the more likely it will be misinterpreted. Second, there are costs that B thinks A could impose but which are not actually threatened. For example, in planning for military action against A, B’s military leadership may assume certain reactions on the part of A and assess the consequences accordingly. Third, there are the costs that actors C, D, E, etc. could impose on B but which are not threatened either but that B is still worried about. Fourth, there may also be domestic costs that are unrelated to the costs imposed by external actors. In each case the deterrent effect on B may have nothing to with A’s threat. Instead, the consequences that actually produce deterrence are those that B imagines will follow from taking a course of action, rather than those threatened, explicitly or implicitly, by A. Thus, as these are consequences that exist in B’s head but may not exist in reality, nor communicated by a different actor, they are more aptly referred to as self-deterrence.
How might this work in practice? Is it possible, for example, for state actors to be deterred, not by fear of military consequences, but due to the fear of undermining their international reputation, or of abstract unintended consequences that may not necessarily be pinned down but still induce caution? When contemplating the prospect of nuclear first use in conflicts against a non-nuclear state, two consequences typically discussed are the reputational damage that would ensue, as well as the ‘letting the nuclear genie out of the bottle’ effect, often conceptualized in terms of other states also using nuclear weapons and/or of more states racing to acquire them. There is considerable evidence of this sort of thinking, such as during the Korean and Vietnam wars, when the possibility of nuclear use was seriously contemplated by senior policymakers before being rejected. In most cases, however, it is assumed that the prospect of nuclear use was rejected by policymakers unconsciously. It would not arise in their correspondence or mentioned in policy meetings with the result no paper trail is left behind.
The prospect of losing one’s reputation is just one of many reasons why state and non-state actors refrain from acting in ways that analysts often assume they will act. How to explain numerous aspects of Russia’s ongoing war in Ukraine, specifically their unwillingness to pursue certain actions, such as not using greater levels of force earlier in the campaign, refusing to call the war a ‘war’, the long delay mobilizing reserves, Russia’s unwillingness to use nuclear weapons for demonstrative purposes or on the battlefield to facilitate a military victory, etc. With each of these we can assume the lack of action was due to reluctance rather than inability. The question is why was the Russian leadership reluctant? What was it they feared? Was it the external consequences threatened, or something else? On the lack of nuclear use, it may be the case that external actors have threatened consequences and that these have been successful. But as we lack information, the jury is still out, and will probably remain so indefinitely. Nevertheless, even if consequences have been threatened by external actors, it is reasonable to assume that other consequences not threatened will also have induced caution.
Consequences such as these that are self-generated are difficult to class in the traditional category of deterrence, which in the mainstream discourse refers to the fear one actor imposes on another. On the other hand, the spectrum of explanations for inaction extends beyond deterrence and self-deterrence. There are many other reasons why states are not constantly at war with one another, why small wars don’t expand into larger wars, and so forth. Moreover, though self-deterrence considerations may be insufficient to convince policymakers not to act, what is important is that they are prominent in policymakers’ discourse when decisions about going to war are taken and they can also shape the way states go to war.
In contrast to Sergey’s call to drop ‘self-deterrence’ from the IR discourse on the grounds that it is ‘completely nonsensical’, I would argue it needs to be embraced by IR scholars and properly studied. I’ll even go one step further and suggest that in many cases, self-deterrence provides a better explanation than deterrence for why states avoid taking certain courses of action in international crises and wars. Regardless though of which has more explanatory power in any specific case, the fact remains that an objective appreciation of a state’s fear of consequences must account for a wider range of consequences than are normally recognized as relevant. Given the multitude of deterrence failures that can be traced back to overly narrow assumptions about an adversary’s cost-benefit calculus, policymakers and analysts tasked with warning about and preventing future deterrence failures would do well to place greater emphasis on self-deterrence rather than throw it in the dustbin.