The White House’s reported retirement of a set of policy restrictions on the use of lethal action against terrorist targets should bring into focus what constraints should apply—as a matter of law—to targeting operations that take place away from active war zones. According to public accounts, the previous targeting restrictions—the Presidential Policy Guidance—imposed a policy overlay that required US forces to meet a much higher standard than the laws of war require. That standard assessment is, however, grounded in a misunderstanding. Legal obligations under the law of war, on their own, may approach the level of certainty required by the policy guidance in the same situations in which the policy itself applied in certain circumstances, including when using force away from States that are commonly recognized as active war zones. In other words, in certain contexts the law of armed conflict may require a near certainty that the proper target is being engaged and that civilian casualties are minimized. If the White House now allows the Defense Department to peel back the “near certainty” standard, it will be the responsibility of U.S. forces to once again maintain the baseline imposed by international law.
We should begin with an important caveat: Our analysis may not be as relevant to battlefields in which the scale and tempo of armed violence escalates to a high level. However, news reports that have proven reliable in the past suggest that the White House may expand U.S. operations into new countries and new theaters which may be remote from hot battlefields.
We feel safe in assuming that the new playbook for lethal actions against terrorist targets includes a savings clause, meaning that the policy prescriptions may not be interpreted to drop below the level required by international law. So, what is the floor imposed by international law? The answer is that it varies according to conditions on the ground, and areas outside active hostilities may impose greater restrictions.
I. The targeting questions at issue
Two targeting decisions are relevant to our analysis.
1. Direct targeting decision: ensuring that the object of attack is a legitimate military target (e.g., an enemy combatant) and not a civilian (principle of distinction)
2. Indirect targeting decision: ensuring that the incidental loss of civilian life is not excessive in relation to the expected military benefit (principle of proportionality)
The Trump administration’s policy guidance reportedly includes the following two constraints:
Implicit precondition: Determining an individual is a terrorist target
Explicit condition 1: Reasonable certainty that the specific, approved terrorist target (i.e., a particular individual, group or object) is present (the Obama-era policy required “near certainty”)
Explicit condition 2: Near certainty that civilians will not be injured or killed.
The implicit precondition highlights an element that is otherwise not clearly included in this policy framework. The Obama-era playbook, for example, included its version of Condition 1—a level of certainty that the terrorist target is present—but did not provide the standard for the antecedent determination that an individual is a terrorist target within the Presidential Policy Guidance. As discussed below, that standard was set forth in subordinate military rules of engagement and targeting guidance. The precondition in the guidance highlights that prior question, which must be settled before making a determination that the other two conditions are met.
There are two ways to read Condition 2. The first reading is that it is a proportionality rule, addressed only after determining that an individual is a terrorist target and is present. In other words, these conditions operate like a series of boxes to be checked off, and Condition 2 is checked off after answering the distinction question (Condition 1). On an alternative reading, Condition 2 is a rule of distinction as well as proportionality. A commander, for example, cannot have near certainty non-combatants will not be killed if he or she does not have near certainty that the target is a combatant. Indeed, this second reading would mean there is no need to include the precondition above, since Condition 2 would fully satisfy it.
We should also note another feature of these two conditions and the way they work with the precondition. It is impossible to raise a legal concern under the law of armed conflict with respect to the mere lowering of the standard in Condition 1 from a “near certainty” to a “reasonable certainty.” That’s because the condition, strictly speaking, involves only the question whether the specific individual is present, not whether the specific individual is a lawful target. If one determines, for example, that a group of enemy fighters (all of whom are lawful targets) are in a location and that the specific individual is, let’s say, somewhat likely to be present (or even unlikely but possible), nothing in the law of armed conflict would forbid striking the group. The entire group is targetable, and it really is just a policy choice whether to restrict attacks only to when the specific individual is present (at whatever level of certainty or uncertainty).
Our analysis so far raises some important side questions. Does the Trump administration consider Condition 2 strictly a proportionality rule or also a distinction rule? How does that understanding get implemented in targeting operations? Does the Trump administration have a standard of certainty that applies to the precondition, and does it vary according to the situation of the decision-maker—depending, for example, on whether a decision-maker is on a hot battlefield making urgent choices or in an air conditioned office in Washington DC authorizing an operation against an identified individual far in advance—or perhaps according to the value of the target and a perceived urgency in striking?
It is also important to consider potential modifications of these standards of proof over time. The current policy framework reportedly retains the ability for the President to suspend or modify any of the restrictions for a given operation or in a specific area. The President would retain such legal authority in any case because this is merely presidential policy guidance. The framework operates, one might say, at the pleasure of the President. It is significant though that the policy framework reportedly includes the option for such variances within it. That design feature means exercising the authority to make exceptions is understood as an application of the policy framework, not a revision or overriding of the framework.
Critics may claim this structural design makes variances easier to adopt and to socialize within the administration, and that such divergence from the baseline restrictions should be more difficult to adopt. It is generally more onerous to alter Presidential-level guidance, directives, or executive orders. Such an action may entail prior interagency review, for example. Having exceptions baked into the guidance from the start avoids transaction costs and other internal executive branch procedures that may be required to override the framework each time. Critics may not like that type of efficiency.
Critics may also claim that including such waivers or downgrades within the framework itself allows the administration to hide from the public when an operation does not abide by the set of default rules. The administration’s spokespersons may simply be able to say that a strike took place in accordance with the policy guidance, without saying whether any specific variances applied.
Regardless of these concerns, and they are genuine ones, the ready availability of variances provides an important reason for our discussion: identifying the floor imposed by international law, below which no variance can lawfully drop. A key question, for example, is whether a downgrade from a near certainty standard to a reasonable certainty standard for Condition 2 is always legally permissible in the locations where the playbook applies.
II. Application of “Reasonable Certainty”—A sliding scale
In important circumstances, the law of armed conflict may require an attacker to obtain a high level of certainty that a target is a legitimate military object (e.g., a combatant). The governing standard for all targeting decisions is one of reasonableness, and it would be unreasonable for a decision-maker to forego such a level of certainty in some cases. A heightened level of certainty should be applied, for example, when conditions approach the end of the spectrum in which high-level commanders or decision-makers have the luxury of significant intelligence information and time to determine whether to authorize a strike several days, if not weeks, in advance; have multiple opportunities to strike the target without any threat to their own personnel and little if any known threat to others; and local conditions are far from an active battlefield. Once one acknowledges that such extreme situations on one end of the spectrum require a higher level of certainty, it is easy to understand how the entire analysis involves a sliding scale. The greater degree to which conditions approach that end of the spectrum, the greater standard of proof required.
It is important to set some nomenclature aside. The term “near certainty” does not have a time-honored usage, and thus is not widely and readily understood; nor is it necessary for a legal analysis. The term “reasonable certainty,” however, has operated in U.S. and other foreign military practice. It is a concept that is capacious enough to require a standard of proof like clear and convincing evidence or a standard more closely approximating beyond a reasonable doubt. The standard is informed by international law, but international law fails to provide a uniform targeting standard to be applied across conflicts. Rather, what is required to obtain reasonable certainty will differ according to facts on the ground.
The concept of reasonable certainty has been a part of U.S. targeting operations. The Joint Chiefs of Staff’s Collateral Damage Methodology, declassified in 2009, defines positive identification of a target as “the reasonable certainty that a functionally and geospatially defined object of attack is a legitimate military target in accordance with the Law of War and applicable ROE.” Israeli armed forces also require “reasonable certainty” in targeting decisions involving the principle of distinction. Holding itself out as an adherent to international law, Israel explained that the law required “reasonable certainty” during the 2014 Gaza Conflict:
“Despite the serious challenges posed by the conduct of Hamas and other terrorist organisations in the Gaza Strip, and despite the complex nature of urban warfare, the IDF remains committed to the Law of Armed Conflict. Accordingly, in the complex circumstances of the 2014 Gaza Conflict, Israel undertook to attack objects only when there was reasonable certainty — based on reliable intelligence — that they constituted military objectives in accordance with the Law of Armed Conflict. Likewise, the IDF directed attacks on individuals only when there was reasonable certainty that such persons were members of organised armed groups or civilians directly participating in hostilities.”
Lt. Col. John Merriam explained the sliding scale involved in the application of “reasonable certainty” in targeting operations:
“This problem is only compounded by the definition that follows, which begins with “a reasonable certainty.” This formulation does include the critical word “reasonable.” However, it does not use “reasonable” in the right way; by linking it to certainty, it takes on the character of a quantum. One is left with the impression that it denotes a level of certainty—perhaps short of “absolute certainty,” but “certainty” nonetheless. As the preceding examination of the law of distinction shows, the concept of reasonableness does not amount to certainty, or to any specific quantum of evidence. Rather, it is a flexible standard that is amenable to use under the full range of conditions that may be present in war. At a high echelon of command, with ample intelligence, surveillance resources, and time to refine the information presented, reasonableness may very well approach certainty. But at the ground level where an individual combatant or a small-unit commander must make decisions under compressed timelines and with perhaps far fewer sources of information, reasonableness requires much less.” (emphasis added)
A heightened level of certainty is easily grounded, as Adil Haque explains in a Just Security essay, in two foundational rules in the law of armed conflict—the obligation on attackers to presume that an individual is a civilian in case of doubt and the more general obligation to take all feasible (or “reasonable”) precautions to ensure that the object of attack is a legitimate military target. Law of war treaties define “feasible precautions” as “those precautions which are practicable or practically possible taking into account all circumstances ruling at the time.” It is easy to understand that a heightened level of certainty will be “practicable or practically possible” under some conditions.
The obligation to take feasible precautions applies to the principle of proportionality, not just the principle of distinction. After selecting a legitimate military object, commanders must take all reasonable steps both to minimize the loss of civilian life and to ensure the loss of civilian life is not excessive in relation to a concrete and direct military advantage. The elements described by Lt. Col. Merriam in his exposition of the sliding scale apply here as well: when it comes to a high echelon of command with significant intelligence information and time to refine that information before taking a strike, it will be more feasible—i.e., “practicable or practically possible”—to obtain a higher level of certainty. Similarly, at a greater level of abstraction, the greater degree to which a targeting operation occurs outside of a geographic region designated as an area of active hostilities, it will likely be more feasible to obtain a higher degree of certainty before authorizing a strike.
Geoffrey Corn has engaged in a similar exercise. His scholarly effort is to provide very specific standards of proof for different situations, which are defined by the types of factors we have described to explain the sliding scale and that Merriam sets forth in his analysis. Corn’s proposal aligns closely with our sliding scale. He writes, for example, that the “quantum of information” necessary to render a decision to attack reasonable against an unconventional enemy outside an area of active combat operations should be very high, and he explains that a definite standard could facilitate the application of the reasonableness test.
Finally, our explanation is consistent with the position of the International Committee of the Red Cross, which explains that the law requires targeting decisions “must reflect the level of certainty that can reasonably be achieved in the circumstances” and that “in practice, this determination will have to take into account, inter alia, the intelligence available to the decision maker, the urgency of the situation, and the harm likely to result to the operating forces or to persons and objects protected against direct attack from an erroneous decision.”
In sum, it does not matter as much what lexical terms the U.S. military adopts in the application of its targeting rules as a matter of policy. To comply with the law of armed conflict, for decision-makers who are operating “at a high echelon of command, with ample intelligence, surveillance resources, and time to refine the information presented, reasonableness may very well approach certainty.” That is true whether or not you call the policy standard “reasonable certainty.”