In a previous post, I argued that a recent speech by Jeremy Wright QC MP, the Attorney General of the United Kingdom, left quite unclear whether the lawfulness of anticipatory self-defense is supposed to turn on the imminence of an armed attack, the immediate necessity of defensive action, or some combination of the two. In this post, I will discuss another respect in which the Attorney General could have done more to clarify his government’s legal position. Since Brian Egan, the US State Department’s Legal Adviser, took a similar approach in an important speech of his own, my concerns apply to the US legal position as well.

Ideally, the Attorney General would distinguish between (1) substantive legal rules; (2) evidentiary standards; and (3) reliable evidence. In other words, he would pose three distinct questions: First, under what factual circumstances does a State have a legal right to use defensive force? Second, given that States are not omniscient, when is the available evidence that these factual circumstances exist sufficiently strong as to warrant the use of defensive force? Third, what constitutes reliable evidence that these factual circumstances exist?

Instead, the UK Attorney General and the US State Department endorse a non-exhaustive list of “factors to consider in asking whether or not an armed attack . . . is imminent,” drawn from a much-discussed article by Sir Daniel Bethlehem:

(a) The nature and immediacy of the threat;
(b) The probability of an attack;
(c) Whether the anticipated attack is part of a concerted pattern of continuing armed activity;
(d) The likely scale of the attack and the injury, loss or damage likely to result therefrom in the absence of mitigating action; and
(e) The likelihood that there will be other opportunities to undertake effective action in self-defense that may be expected to cause less serious collateral injury, loss or damage.

As Monica Hakimi observes over on EJIL:Talk!, “[i]t is not clear how these factors relate to one another, or how much weight any particular one carries.” I would argue that these factors are different in kind, not merely in weight, and are relevant to distinct legal questions. Several of my concerns overlap with those raised several years ago by Dapo Akande and Thomas Liefländer, and which the Attorney General would have done well to address.

In my view, the nature and immediacy of the threat are best understood as part of the substantive legal rule proposed by the UK—that States may use defensive force to prevent an imminent armed attack (or, perhaps, when immediately necessary to prevent a more temporally remote armed attack).

In contrast, the probability of an attack alludes to an applicable evidentiary standard or, in Bethlehem’s terms, whether there is “a reasonable and objective basis for concluding that an armed attack is imminent.” Indeed, the probability of an attack is just another name for the weight of the evidence that the attack will occur unless prevented. Unfortunately, the Attorney General does not tell us what the applicable evidentiary standard might be, that is, how probable an attack must appear given the available evidence before defensive force is warranted. Instead, he simply identifies “the probability of an attack” as one factor among many.

The Attorney General does say that “it is crucial that we make these assessments ‘in good faith and on the basis of sound evidence’ in order to have a sufficient level of confidence to justify action.” However, he does not tell us what level of confidence is sufficient to justify action, or how that level of confidence might vary based on the circumstances.

Moving on, whether the anticipated attack is part of a concerted pattern of continuing armed activity may be reliable evidence, or corroborate other reliable evidence, that the anticipated attack is imminent or will occur unless immediate defensive action is taken. Again, this factor is different in kind from those before it.

Arguably, the likely scale and effects of an anticipated armed attack may be relevant to the applicable evidentiary standard (as well as to the proportionality constraint, of course). The level of evidence that an armed attack is imminent that is sufficient to warrant defensive force may depend, to some degree, on the relative risks of error, that is, on the likely scale and effects of the anticipated armed attack and the likely scale and effects of the contemplated defensive force. This is a difficult question that should be squarely addressed on its own terms. Presumably, the UK would not endorse the view that States may use defensive force when there is only a one percent chance (that is, very weak evidence) of an armed attack. Beyond that, we are left to speculate.

Finally, the likelihood that there will be future opportunities to undertake less harmful defensive measures suggests that more harmful defensive measures may violate an entirely different legal constraint. As Bethlehem writes, “[a]rmed action in self-defense must be limited to what is necessary to address an imminent or actual armed attack” (Bethlehem, Principle 3). Accordingly, if it is likely that an armed attack can be effectively repelled or prevented by less harmful defensive force later, then it may be unlawful to use more harmful defensive force now. None of this has much to do with imminence, or so it would seem.

The Attorney General’s failure to identify an applicable evidentiary standard seems most unfortunate. Many States, legal experts, and ordinary citizens who might be willing to entertain the abstract idea of anticipatory self-defense balk precisely because they worry that States will either abuse such a principle or apply it recklessly, on the basis of weak or unreliable evidence. I fear that the Attorney General’s speech will do little to alleviate these grave concerns.