Understanding Self-Defense and the Law of Armed Conflict


| Mar 9, 2021

Self-Defense. Checkpoint in Daab Pass in Shinkai district, Afghanistan Feb. 25. US Army


Many soldiers, commanders, and judge advocates involved in the conflicts in Africa, Syria, Iraq, and Afghanistan deploy with a dated and incomplete understanding of individual self-defense. This reality is the product, in part, of definitional imprecision in the 2005 U.S. Standing Rules of Engagement (SROE), theater and mission-specific Rules of Engagement (ROE), and tactical and operational directives. In our estimation, the SROE’s single standard for targeting individuals on the basis of conduct has fueled an erroneous perception on the part of many that self-defense is the legal justification for employing force during engagements in armed conflict. The SROE’s failure to provide explicit guidance on using force against civilians who directly participate in hostilities adds confusion and risks mission accomplishment.

Recent scholarship has sought to clarify the parameters of the right of individual self-defense in armed conflict (see, for example, here, here, and here), but confusion remains. For instance, at an International Society for Military Law and the Law of War conference for military legal advisors, a vignette concerning soldiers’ use of force against a fast-approaching vehicle at a traffic control point generated an interesting discussion regarding the proper legal basis for the engagement. Some in attendance concluded that self-defense was the proper basis for the use of force. We disagree. In our opinion, the better approach would be to treat the situation as one involving direct participation in hostilities.

In this post, we examine the legal basis for using force against individuals during armed conflict, with emphasis on the notions of individual self-defense and direct participation in hostilities. In particular, where and how does self-defense fit into the law of armed conflict (LOAC)—if at all?

The Law of Self-Defense

Concepts of self-defense reside in two distinct bodies of public international law. One concept appears in the jus ad bellum, or the law applicable to States’ resort to the use of force. Article 51 of the U.N. Charter recognizes the inherent right of States to engage in self-defense in the face of an “armed attack.” This article does not address the jus ad bellum self-defense.

Self-defense is also an implied right derived from an individual’s “right to life” found in numerous international human rights law agreements—most prominently the International Covenant on Civil and Political Rights (Article 6). It is a right that finds expression in domestic criminal law, as is the case with Rules for Court Martial 916 of the Uniform Code of Military Justice and Section 3 of the United Kingdom’s Criminal Law Act of 1967.

With rare exceptions, however, the notion of individual self-defense is conspicuously absent from the law of armed conflict. The four 1949 Geneva Conventions make no direct mention of self-defense. Conventions I (Article 22) and II (Article 35) indirectly raise the issue, but only with regard to defense of medical personnel and the wounded and sick in their charge. Additional Protocol I—a treaty to which the United States is not a party—only mentions self-defense in one provision, Article 67(1)(d). That provision involves the arming of civil defense personnel for self-defense. Additional Protocol II contains no reference to self-defense. And while the ICRC Commentary to the Additional Protocols mentions self-defense (para 560), it does so only in reference to criminal acts committed by “uncontrolled elements or looters” during armed conflict outside the scope of LOAC.

Finally, there is no support for concluding that customary provisions of the law of armed conflict provide any basis for using force in individual self-defense. In fact, the sole mention of self-defense in the ICRC’s 2009 Customary International Humanitarian Law study is in the commentary to Rule 29, which discusses medical personnel using light arms in self-defense.

This does not mean that self-defense is never relevant during armed conflict. Situations without a nexus to the hostilities can arise where soldiers may act in self-defense. For example, a soldier may use force in defense of self or others during occupation in response to a criminal attack that is unrelated to the armed conflict. In this instance, however, the right to use force derives not from the law of armed conflict’s conduct of hostilities rules, but from the law applicable to maintaining order during the occupation (a form of domestic law) and international human rights law.

Targeting Individuals under the Law of Armed Conflict

The absence of reference to self-defense in those LOAC rules begs the question of the legal basis for using force during armed conflicts.

Under the law of armed conflict, soldiers may lawfully engage certain individuals at any time based on their status. Civilians may be targeted based on their conduct, specifically while they are directly participating in hostilities. The conduct that qualifies as direct participation in hostilities is not defined in the relevant treaties that refer to civilians losing their protection from attack—Additional Protocols I and II. And, the exact scope of the term remains the source of disagreement (see here and here).

The U.S. position is set forth in the Department of Defense Law of War Manual and the U.S. Army and U.S. Marines Corps Commander’s Handbook on the Law of Land Warfare (FM 6-27). According to those manuals, “taking a direct part in hostilities extends beyond merely engaging in combat and also includes certain acts that are an integral part of combat operations or that effectively and substantially contribute to an adversary’s ability to conduct or sustain combat operations.” Both manuals confirm that direct participation in hostilities is highly contextual. They provide a list of relevant considerations, including “the degree to which the act causes harm to the opposing party’s person or objects.” Notwithstanding the difficulty of making case-by-case assessments, the bottom line is that civilians who directly participate in hostilities forfeit their protection from attack under the law of armed conflict for such time as they are doing so.

It is also important to stress that regardless of whether the soldiers are in an offensive or defensive posture when engaging civilians directly participating in hostilities, they are not acting in self-defense. To the contrary, they are authorized to engage because the civilians forfeited their protection from attack by taking a direct part in hostilities. This is often where we find confusion among U.S. forces. Soldiers have come to understand “self-defense” as both a legal term and a legal requirement for kinetic engagements. This misperception leads to self-imposed and unnecessary tactical restrictions that could have long-term impact, especially in near-peer conflict. We can reverse this alarming trend by crafting practical rules of engagement focused on concepts of direct participation in hostilities that can be seamlessly operationalized by the soldier.

The Rules of Engagement Challenge

Recent armed conflicts illustrate the challenge in distinguishing those who are targetable based on status or conduct from the civilian population. Rules of engagement help soldiers navigate the complexities of distinction, but these ROE must be clear and effective.

The 2005 SROE—first implemented in 1994 and amended in 2000—is the foundational document for rules of engagement. It provides “implementation guidance on the application of force for mission accomplishment and the exercise of self-defense.” The SROE combines national policy goals, operational guidance, and the law to set forth the policies and procedures governing the actions of the U.S. armed forces during operations outside U.S. territory.

Accounting for the dynamic nature of armed conflict, the SROE provides commanders flexibility by allowing them to request additional authorities to use force or restrict its use, called “supplemental measures,” to accomplish their assigned mission. Using the SROE as the foundational document, Combatant Commands (e.g., U.S. Central Command) and/or Theater Commands (e.g., U.S. Forces Afghanistan) often request supplemental measures to generate “mission accomplishment” ROE that guide their subordinate units at the tactical level.

For instance, mission accomplishment ROE can authorize the use of force against declared “hostile forces.” Hostile forces are “[a]ny civilian, paramilitary or military force, or terrorist(s) … declared hostile by appropriate U.S. authority.” There is no requirement for members of the forces declared hostile to have demonstrated any hostile intent or committed a hostile act. It is rather their status as members of an organized armed group, dissident armed forces, or a State’s armed force that provides the basis for targeting them. This accords to the rules under the law of armed conflict that allow for targeting members of such groups at any time.

The SROE, however, is more confusing when it comes to conduct-based targeting. The SROE  links hostile act/hostile attempt to self-defense when it states, “(u)pon commission of a hostile act or demonstration of hostile intent, U.S. forces may use all necessary means available and all appropriate actions in self-defense.” This guidance may prove useful for commanders and their units outside of armed conflict. However, the SROE fails to provide authority to engage civilians taking a direct part in hostilities during armed conflict.

Consequently, confusion derives from the fact that the SROE, theater and mission specific ROE, as well as tactical and operational directives, sometimes refer to self-defense concepts as they attempt to translate complex legal principles into workable guidelines for those at the tactical level. This reference to self-defense can have the opposite effect by causing commanders, soldiers, and judge advocates to believe that self-defense is the legal justification for their use of force and that the situation must be one of self-defense before engaging an adversary.

As a result, soldiers may hesitate to engage lawful targets under LOAC because they mistakenly understand ROE to require a demonstrated hostile act or hostile intent. For example, self-defense would not permit the lethal engagement of an individual who had been involved in a firefight with friendly forces but had just left the scene of the engagement. However, under a direct participation paradigm, that person could be targeted beyond the actual engagement, to include while leaving the scene, and possibly for a longer period, if the individual were likely to engage in future attacks. As COL Randy Bagwell notes, the SROE has created a “confusing self-defense landscape” for commanders, their troops, and judge advocates.

To illustrate the confusion, consider Operation Iraqi Freedom and the Multi-National Corps-Iraq (MNC-I) ROE card disseminated in 2007 to more than 150,000 Soldiers, Sailors, Airmen, and Marines deployed to Iraq. The ROE card does not address status-based targets, but states:


  1. You may engage the following individuals based on their conduct:
    • Persons who are committing hostile acts against CF [Coalition Forces].
    • Persons who are exhibiting hostile intent towards CF.

While this cautious approach might make sense operationally in counterinsurgency or stability operations, use of terms drawn from SROE self-defense rules is problematic. The card could be misunderstood as indicating that there is no law of armed conflict basis for targeting individuals based on status alone (i.e., declared hostile forces), nor does the card address civilians directly participating in hostilities. It risks teaching—unintentionally—operators and judge advocates the wrong lesson.

There is nothing wrong with an operational commander limiting engagement authority for operational or policy reasons, but this must be done in a manner that does not imply nonexistent international law limitations. ROE must establish authority to use force, as appropriate, in armed conflict and/or situations outside of armed conflict. Limiting ROE to considerations of hostile act and hostile intent dangerously ignores the prospect that soldiers will face civilians who take a direct part in hostilities. Of equal concern, soldiers have been conditioned to think in terms of ROE and self-defense instead of understanding and incorporating the principles and practical considerations of direct participation.

Engaging Individuals Directly Participating in Hostilities

The vehicle vignette mentioned above provides the opportunity to analyze a situation in which soldiers are authorized to target individuals based solely on their conduct, not individual self-defense.

Considering the U.S. position, the driver of the vehicle may be targeted with lethal force as an individual taking a direct part in hostilities. Speeding toward the checkpoint, the driver’s aggressive approach to the obstacles at the traffic control point reasonably indicates he is likely trying to kill or injure the soldiers and damage the checkpoint. With the limited window of opportunity to make such determinations on the battlefield, the actions of the driver could lead a reasonable soldier, in the same or similar circumstance, to conclude that the driver is directly participating in hostilities and may be engaged on that basis. But it is important to emphasize that even though this may look like a self-defense situation, the legal basis for engaging the driver is not a conclusion that the soldier needs to defend himself, but rather that the soldier has concluded the driver is directly participating in hostilities.

Importantly, even if the soldiers were incorrect in their assessment that the driver was taking a direct part in hostilities, it is notable that there would be no violation of LOAC. First, LOAC’s “Rendulic rule” assesses the actions of soldiers through their perspective at the time they were forced to make the decision to engage the driver, not with the benefit of hindsight. Marko Milanovic has suggested that this type of mistake of fact would not be a violation of LOAC, assuming the mistake was both honest and reasonable.

Second, as noted in the DoD Law of War Manual (para, the United States does not consider the presumption of civilian status, otherwise known as the “rule of doubt,” set forth in Articles 50(1) and 52(3) of Additional Protocol I, to be customary in character. Instead, U.S. policy authorizes soldiers in the above scenario to engage the driver if they conclude “in good faith based on the information available to them in the light of the circumstances ruling at the time” that the driver was directly participating in hostilities.

Concluding Point

The legal distinction among status-based targeting, direct participation in hostilities, and self-defense targeting is critical for judge advocates to understand and to impart to commanders and operators. This must be done in a manner that distinguishes operational and policy choices from legal constraints. Failure to do so may create a situation in future conflicts where soldiers hesitate to act when necessary because they believe it is unlawful to do so.

The largest hurdle in this effort may very well be the mindset of our soldiers and the oft-repeated refrain that they “engaged the enemy in self-defense.” Our troops have internalized self-defense as the authority to use force in armed conflict, a reality that is counter to LOAC and the operational environment. Commanders, with the assistance of their judge advocates, must work to reframe the ROE mindset by incorporating practical direct participation in hostilities guidance that reflects the reality of combat operations and prepares our warfighters for today’s combat operations and also the potential for near-peer conflict in the future.


LtCol John R. Cherry is the Deputy Chair and Military Professor at the Stockton Center for International Law at the U.S. Naval War College (NWC) in Newport, RI.

LTC Michael Rizzotti is an Army Judge Advocate currently serving as the Chief, Complex & Capital Litigation, U.S. Army Trial Defense Service at Fort Belvoir, Virginia.

The views expressed here are their own and do not represent those of the U.S. Naval War College, the Department of the Navy, the Department of Defense, or any part of the U.S. government.