[Editor’s Note: On Tuesday, the Department of Defense released an updated version of its Law of War Manual. Just Security is hosting an online “mini forum” to discuss the new changes to the Manual.]

The important revision to the Law of War Manual released today will undoubtedly garner substantial scrutiny and comment. Having been provided the opportunity to review these provisions just prior to release, I hope to help stir discussion on these revisions. I will therefore focus on what I consider to be the three revisions I consider most positive, and the three areas I think missed the mark.

As an initial matter, I share the view that DoD should be applauded for engaging in this revision process. As most readers know, DoD has been remarkably open to encouraging both internal and external discourse on the Manual. Many readers have been engaged in this discourse, and many wondered if and how the General Counsel would respond to criticisms and suggestions. This (and the prior) revision indicates the DoD’s genuine commitment to continually improve this Manual. That is both laudable and important. This Manual has and will continue to have a profound impact on the planning, execution, and critique of U.S. military operations, and those who rely on it deserve a commitment to periodic review, update, and improvement. The Manual’s impact on the evolution and interpretation of the law of armed conflict, and the practices of other armed forces, reinforces the significance of the commitment to constant improvement. The importance of this improvement process cannot, therefore, be underestimated.  

Three Positive Developments

First, the emphasis on the relationship between “targeting law” with operational practice.

Paragraph 5.10.3 addresses the importance of targeting process in the civilian risk mitigation equation, a point emphasized by the General Counsel in her recent speech at New York University. The Manual notes that procedures contribute to implementation of the proportionality obligation, and includes as examples of such procedures, “rules of engagement, doctrine, standard operating procedures, and special instructions.” The paragraph then notes that, “These procedures help ensure the sound implementation of the requirements of the principle of proportionality.”

This paragraph reflects something that is probably fairly characterized as axiomatic for military practitioners, but nonetheless necessitated emphasis. As I noted in this article co-authored with my brother Colonel Gary Corn, targeting law is and must be aligned with operational practice, and is best understood through that lens. The process used to implement that law integrates the law into the operational targeting process, and in so doing enhances the civilian risk mitigation effect the law seeks to achieve.

The importance of this integration is also reflected in paragraph, which emphasizes the symmetry between targeting law and the operational principle of economy of force, noting that, “Use of excessive and indiscriminate force is costly, highly inefficient, and wasteful of scarce resources.” In so doing, commanders and legal advisors are reminded that compliance with proportionality – or more specifically foregoing attacks based on the assessment that the anticipated civilian harm will be excessive in comparison to the anticipated military advantage – will often be consistent with the operational logic of carefully allocating combat power only where it will produce a meaningful operational payoff. As I understand it, the Manual is indicating that when the military value of attacking a target is outweighed by the risk of excessive civilian harm, that value does not warrant the expenditure of finite combat resources. This might not be intuitive for commanders, and thus the emphasis is potentially beneficial. Indeed, as explained in the article referenced above, I believe there is important value in emphasizing the alignment between LOAC targeting law and other operational principles, such as objective, mass, and of course legitimacy, and wondered why economy of force was the only principle invoked?

I think this emphasis on the symmetry between operational practice and LOAC targeting law generally, and the significant role targeting process plays in advancing the objectives of this law, is vital and long overdue. Indeed, this was the thesis of my 2015 article titled, War, Law, and the Oft Overlooked Value of Process as a Precautionary Measure. As the Manual notes, while LOAC targeting rules apply at every level of military command and during every aspect of a military operation, their impact is most significant during the deliberate targeting process (paragraph 5.10.4, “Thus, although the legal standard of proportionality does not distinguish as such between, e.g., dynamic and deliberate targeting, the availability of more time in the context of deliberate targeting generally allows for more actions to be taken to reduce the risk of civilian casualties.). Accordingly, the positive effect of the law will inevitably be enhanced by an increased understanding of the relationship between the process and substance of target decision-making. The Manual’s clear emphasis on the importance of this interrelationship is therefore a major advance in effective LOAC implementation.

Second, elevating the significance of precautionary measures in the process of assessing target legality

As noted in the precautions article referenced above, and in other articles I have written, I have long believed precautionary measures have been undervalued in our LOAC discourse, education, and implementation (a view shared by a number of highly respected colleagues, for example as expressed by Brigadier General (Retired) Kenneth Watkin in, Military Advantage: A Matter of Value, Strategy and Tactics, 17 YBIHL 277, 311-14 (2014)). While debates over anticipated military advantage and the balance between advantage and civilian risk tend to dominate this discourse, I believe that the more concrete process of working through feasible precautions is more aligned with operational and tactical logic, and is in fact the most common way in which military commanders and their subordinates endeavor to mitigate civilian risk. I have also argued that effective implementation of the precautions obligation will often reduce the complexity of the ultimate proportionality assessment by reducing civilian risk before that assessment must be made.

The Manual now seems to fully embrace this view by more directly folding considerations of feasible precautions into the enumeration of proportionality obligations. Specifically, the first paragraph of the revision, Proportionality in Conducting Attacks, provides that,

In accordance with the principle of proportionality, combatants must not exercise the right to engage in attacks against military objectives in an unreasonable or excessive way. Therefore, when prosecuting attacks against military objectives (i.e. , the persons and objects that may be made the object of attack), combatants must exercise due regard to reduce the risk of incidental harm to the civilian population and other persons and objects that may not be made the object of attack. In particular, the following rules apply:

  • Combatants must take feasible precautions in planning and conducting attacks to reduce the risk of harm to civilians and other persons and objects protected from being made the object of attack; and
  • Combatants must refrain from attacks in which the expected loss of civilian life, injury to civilians, and damage to civilian objects incidental to the attack would be excessive in relation to the concrete and direct military advantage expected to be gained.

Paragraph 5.10.5 then explains that,

This manual describes the requirement to take feasible precautions in planning and conducting attacks and the prohibition on attacks expected to cause excessive incidental harm as rules that are based on the principle of proportionality. Other sources may describe the relationships between the rules differently. But this manual adopts this approach because the requirement to take feasible precautions in planning and conducting attacks and the prohibition on attacks expected to cause excessive incidental harm are fundamentally connected and mutually reinforcing obligations.

The prior revision did indicate that precautionary measures are a subset of proportionality assessments. However, that revision was not nearly as emphatic. This more direct linkage between precautions and proportionality will ideally enhance U.S. implementation of LOAC targeting obligations. It also reinforces an important point: the implementation of precautions, or failure to do so when feasible, can be highly probative when assessing compliance with or violation of the proportionality obligation. Commanders (and the lawyers who advise them) should understand that their efforts to implement precautions is an important indication of commitment to their overall obligation to mitigate civilian risk, and may be significant in informing assessments of compliance with the proportionality obligation.

The Manual gets it exactly right: precautions and proportionality are indeed “mutually reinforcing” obligations. Some may be critical of the conflation of proportionality and precautions, and advocate a more sequential understanding of these principles. But I doubt that the Manual is really advocating anything less, and is instead emphasizing that commanders must constantly consider feasible precautions as a first step in the civilian casualty risk mitigation effort that culminates with proportionality assessments.

One thing seems certain: commanders should be taught from inception that the LOAC expects them – or more accurately demands of them – that they constantly endeavor to utilize measures to mitigate civilian risk so long as doing so will not compromise the effectiveness of their military operations. The Manual’s revised emphasis on the role of precautions, an emphasis also reflected on the General Counsel’s recent speech, will hopefully contribute to cultivating this instinct in the operational force.

Third, reinforcing the legally difficult but operationally logical need to mitigate risk to military medical activities.

Operational lawyers have probably always assumed that a certain degree of “extra” caution is required when an attack will jeopardize military medical facilities, personnel, and the wounded and sick (essentially, anything or anyone bearing the protective emblem). However, it is far from clear that the law imposes such an obligation, and more specifically whether incidental risk to military medical personnel, facilities, and the wounded and sick must be included in proportionality assessments.

The 2015 Manual answered this proportionality question with an emphatic “no”; the recent revision to the International Red Cross Commentary to the First Geneva Convention answered it with a “yes.” This dichotomy of views produced important discourse and consideration of the applicability of proportionality to military medical personnel and facilities, and to the overall issue of protecting such personnel and facilities from incidental injury and/or collateral damage.

The Manual revision once again wades into this murky water, reinforcing the approach adopted in 2015. That version did indicate that the precautions obligation applied not only to civilians and civilian objects, but to other protected persons and objects, and then listed military medical units and facilities as examples (paragraph 5.11.1). However, in that version, the exclusion of these personnel and facilities from the scope of the proportionality rule did not expressly emphasize the importance of nonetheless applying precautions.

The revised text adopts an obviously different approach. Paragraph begins by reiterating the exclusion of such personnel and facilities from proportionality assessments, noting that,  

The prohibition on attacks expected to cause excessive incidental harm requires consideration of civilians and civilian objects, but this prohibition generally does not require consideration of military personnel and objects, even if they may not be made the object of attack, such as military medical personnel, the military wounded and sick, and military medical facilities.

However, the paragraph then immediately emphasizes the accordant obligation to nonetheless consider feasible precautions to mitigate the risk of inflicting incidental harm on these personnel and facilities,

The exclusion of protected military personnel and military medical facilities from this prohibition reflects such factors as, among others, the general impracticality of prohibiting attacks on this basis during combat operations. . . .

Nonetheless, feasible precautions must be taken to reduce the risk of harm to military personnel and objects that are protected from being made the object of attack. For example, in the context of a deliberate, planned bombardment of a military objective near an identifiable military hospital, it may be feasible to take precautions to reduce the risk of harming the military hospital, and such precautions must be taken.

This is important. By emphasizing a direct link between the proportionality exclusion and the precautions obligation, it mitigates the risk of confusion related to application of targeting “protections” to military medical facilities and personnel.

Still, like the ICRC approach, the extension of the precautions obligation to protected military personnel and facilities raises questions. Unlike the ICRC, the Manual does not invoke the “respect and protect” obligation as the basis for this extension. Instead, like the 2015 version, the revision generally sidesteps identifying the source of this obligation, an issue central to the criticism of the ICRC approach (the 2015 paragraph cited Article 27 of Hague IV, a rule applicable to “sieges and bombardments,” to support of the extension of precautions beyond civilians and civilian objects). In that sense, it is potentially open to analogous criticism: extending a rule explicitly adopted to protect civilians (for the ICRC, proportionality; for DoD, precautions) to protect military personnel and facilities. Nonetheless, the revision reinforces the important rule that protection for these personnel and facilities extends beyond the prohibition against deliberate attack and the pre-attack warning obligation. In my view, extension of the precautions obligation is a more logical approach for establishing of how added protection from incidental harm is best achieved. Because a precautions approach turns on the feasibility assessment, it seems more aligned with operational practice, and candidly less confusing than requiring proportionality assessments whenever an attack might produce such incidental harm (see here for a forthcoming essay on this exact topic).

Some will likely be critical that DoD persists in its position that proportionality does not protect such personnel and facilities. Not me. While I was critical of the original position, which I felt was too dismissive of a fundamental humanitarian dictate (see the same article cited above), I don’t believe that the respect and protect obligation supports the extension of proportionality to any military facility or individual, even the wounded and sick. But I also believe DoD has now reinforced a fair and credible balance of interests: these facilities and individuals may not fall within the protective scope of proportionality assessments, but basic dictates of humanity demand that the prohibition against deliberate attack be complemented by measures to avoid subjecting them to even incidental harm whenever such measures will not compromise the effectiveness of attacking lawful proximate targets.

As noted above, I suspect this compromise approach is actually perfectly aligned with the instincts and practices of commanders and their subordinates. Indeed, I would be dismayed to learn that a commander, knowing an attack may jeopardize military medical personnel or facilities, would be completely indifferent to this risk. But including this qualification to the exclusion of such personnel and facilities from the proportionality obligation is important to reinforce this instinct, and to arm legal advisors with a solid statement of U.S. doctrine that prohibits such indifference. My one regret is that the revision did not include a clear reminder that civilian medical facilities and personnel are always protected not only against deliberate attack, but also by both the precautions and proportionality rules.

Three Points of Concern

First, a Subtle Dilution of the Obligation to Disobey Clearly Unlawful Orders?

One of the most interesting aspects of this revision is the substantial treatment of the process of target selection and engagement, and the relationship between the commander ordering an attack and the subordinates executing that order. I think this is quite beneficial, and reinforces the complexities of combat action, the justifiable presumptive legality of superior orders and accordant expectation that subordinates will normally decisively execute superior orders, and the reality that complex balancing of interests made at one level of command may often be beyond the situational competence of subordinate commands ordered to execute attacks.

However, none of this is absolute, and as the Manual notes, there may be times when a subordinate must question, and perhaps even disobey an order to conduct an attack. In an apparent effort to highlight the complexity of assessing if and when an attack order clearly violates the prohibition against indiscriminate attacks, paragraph specifically addresses the relationship between the duty to disobey and the proportionality rule,

In the context of the principle of distinction, it would often be clear whether a given situation implicates the duty not to comply with clearly illegal orders to commit law of war violations – such as the duty of a subordinate to refrain from complying with an order to attack the civilian population. However, the nuances involved in applying the principle of proportionality could make it more difficult to know whether an order given is clearly illegal.

The duty not to comply with orders that are clearly illegal also applies to violations of the principle of proportionality, in particular, the case of a commander who orders subordinates to conduct an attack that is expected to result in civilian casualties that the commander himself or herself acknowledges would be excessive.

Nonetheless, this duty must be understood in light of application of the principle that law of war obligations are implemented by those with responsibility to make the decisions and judgments required by the law of war, and in particular the point that subordinates might not be competent to assess whether the requirements of proportionality had been met.

Assessing a potential distinction violation is probably much easier than assessing a proportionality violation (at least in the context of an armed conflict against a uniformed enemy who complies with his own “passive” distinction obligation). In that regard, I think the paragraph makes sense. But it is dangerous to suggest that the obligation is somehow qualified by the complexity of the rule itself. Certainly, when read in context the paragraph cannot be understood as a “waiver” of the disobedience obligation in relation to an attack that appears to violate the proportionality obligation. However, might it create a risk of too much reliance on superior attack judgments? Drawing this line is obviously extremely complex. I’m not sure, however, what value was added by the last paragraph.

It is worth noting that this emphasis on the difficult intersection of proportionality judgments, attack orders, and the subordinate obligation to obey all but clearly unlawful orders can cut both ways. In fact, I suspect in the past 15 years there have been far more instances of subordinates chafing under proportionality-based orders that restrict the use of combat power than instances where subordinates felt compelled to disobey an order because they concluded it was too indifferent towards civilian risk. And this is a vitally important point: the limits on a subordinates ability to appreciate complex proportionality judgments made by superior commanders may not only result in executing attacks that may seem to manifest indifference towards civilians, but probably more frequently foregoing attacks that seem completely justified at the tactical point of attack but are incompatible with law and policy at from a more expansive command perspective.

It is also interesting to consider this paragraph in relation to paragraph 5.11.4, Cancelling or Suspending Attacks Based on New Information Raising Concerns of Expected Civilian Casualties, which provides,

However, subordinate commanders or engagement authorities would be required to report promptly new information of expected civilian casualties or to cancel or suspend attacks in appropriate circumstances. For example, aircrews have refrained from conducting an attack or have redirected a munition toward a different impact location when concerns about civilian harm have become unexpectedly apparent.

Subordinates should also generally be understood to have the authority to make decisions required by the law of war, such as the decision to cancel or suspend an attack in light of new information, in order to effectuate the commander’s intent. This is especially true when those commanders will not be able to maintain situational awareness of the risks of civilian casualties.

I don’t think subordinates should, “generally be understood to have authority to make decisions required by the law of war”; I think subordinates have an absolute obligation to make decisions required by the law of war, including the decision to disobey an order they believe is clearly unlawful. This is reflected in the text of Additional Protocol I (AP I), which imposes an obligation on subordinates to suspend or cancel an attack if they determine that executing the attack will result in a proportionality violation that was not contemplated by the superior who ordered the attack. Indeed, while I understand the U.S. concerns related to acknowledging the binding effect of AP I rules, I think invoking it in relation to this issue would have produced needed clarity. I recognize that this is never an easy decision, and is a rare occurrence. But, I also think some of the word choice included in these provisions risks signaling a more qualified obligation than was intended.

Second, an Under-Emphasized Precautionary Measure?

As noted above, I think one of the most beneficial aspects of this revision is how it elevates the significance of precautionary measures in the target legality and civilian risk mitigation process. However, when I reached paragraph 5.11’s list of feasible precautions, I was surprised that the list did not include – actually that it did not include as the first entry – efforts to verify the true nature of a nominated target.

Information and the resulting situational awareness are essential components to ensuring compliance with both the distinction and proportionality obligations. As a result, collecting, assessing, and constantly updating target related information is an essential precautionary measure. In fact, when an avoidable information deficit results in attacking a target that is erroneously assessed to qualify as a military objective, or attacking at target based on a civilian risk calculation that is erroneously too low, it is uncertain whether the attack actually violates distinction or proportionality. What is clear is that where the failure to gather and assess reasonably available information is the primary cause for such an attack, it is the precautions failure that is most obvious and most troubling.

As a result, I had expected to see “information gathering” or “situational awareness” as the first precautionary measure. The importance of information is emphasized in subsequent paragraphs addressing assessment of civilian risk and the role such assessments play in implementing the proportionality obligation. I suspect the drafter’s of the revision would therefore respond that this sufficiently addresses the role of information as a precaution, and that a commander’s instinctive interest in maximizing situational awareness negates the need to specifically enumerate this precaution alongside other precautions such as pre-attack warnings and timing of attack. This may be true, but why not enumerate information gathering as a specific precaution? If we assume that our armed forces seek to implement their LOAC obligations in good faith, we should constantly emphasize that their efforts will always be enhanced by more and better information. If emphasizing this does nothing more than reinforce the commander’s instinct to constantly demand more and better information, it can only enhance LOAC compliance.

Third, a Missed Opportunity to Emphasize the Standard of “Effective” for Advanced Warnings

One of the precautionary measures that is enumerated is the obligation to provide effective pre-attack warnings, “Unless circumstances do not permit, effective advance warning must be given of an attack that may affect the civilian population.” Paragraph then explains “effective” advance warnings by providing,

The purpose of a warning is to facilitate the protection of the civilian population so that civilians and the authorities in control of the civilian population can take measures to reduce the risk that civilians will be harmed by military operations. Although there is no set form for warnings, a warning should be designed to accomplish this purpose to the extent feasible.

I think this was a missed opportunity to emphasize that the effectiveness of an advanced warning must be based on a pre-warning assessment, and not on whether the warning actually produced the intended protective effect. By indicating that, “warnings should be designed to . . .”, the explanation does suggest such an approach to assessing effectiveness, but it is insufficient. Commanders should be encouraged to implement use any method of warning he or she reasonably assesses will be effective, and should not be criticized or condemned if the actual result of the warning does not produce the intended effect.

This may seem self-evident, but criticism of the “knock on the roof” warning tactic suggest otherwise; that some may interpret the term “effective” based on outcome as opposed to reasonably expectation. This should be resisted, because if this takes hold as the standard for assessing compliance, it will stifle creativity and lead commanders to consider only a small range of previously validated warnings. A simple addition to the paragraph indicating that compliance with the warning obligation – both in terms of when circumstances do not permit and whether the warning complied with the effectiveness requirement – must be assessed like any other targeting judgment: was the decision reasonable based on the information available at the time.


I have no doubt that others will identify more “ups” and “downs” with this revision. I also have no doubt that the DoD General Counsel will continue to be receptive of critiques in the ongoing effort to make this Manual more comprehensive, credible, and operationally utilitarian. For me, the strengths of this revision far outweigh any weaknesses, and I believe it will be well-received by the operational forces it is intended to inform. While scholars, observers, and others will form their own opinions on the relative merits of this revision, it is that audience that ultimately matters most, and it is their interests that will hopefully inspire future positive revisions to this important Manual.

Image: Wikimedia