Guest Post: Henderson & Cavanagh on Self-Defense & The Privilege of Combatancy

by Jens David Ohlin

[Jens David Ohlin is Professor of Law at Cornell Law School.]

Ian Henderson and Bryan Cavanagh have hit the nail on the head in identifying a crucial and under-theorized question that goes right to the basic structure of the laws of war. I am in complete agreement that invocations of self-defense during armed conflict are both confused and confusing.

There is already the frequent problem of conflating individual self-defense (in the criminal law sense) with collective self-defense (jus ad bellum and article 51 of the UN Charter).  In addition, one often hears talk of a soldier’s right of self-defense – a claim that is mostly redundant since soldiers hold the privilege of combatancy and have no need to invoke a separate justification for their behavior. A justification like self-defense serves to negate the wrongdoing of the act, but a privileged soldier who kills a legitimate target has committed no wrongful act that requires negating.  Consequently, the justification of self-defense is only relevant during armed conflict when the defender is unprivileged, such as a civilian who has no right to participate in armed conflict.  If the civilian is challenged by an enemy combatant who refuses to obey the principle of distinction, then the civilian is entitled to kill the soldier in self-defense. If a soldier is attacked by a civilian, the soldier can legitimately kill the civilian because he is directly participating in hostilities; no claim of self-defense is required because the privilege applies.

Understanding self-defense from a civilian’s perspective is more difficult.  Consider the complicated problem of a civilian who kills an enemy soldier who is in the process of killing the civilian as collateral damage during a lawful attack against a military objective. In that scenario, does the civilian have the right of self-defense? In the criminal law we usually view the right of self-defense as only applicable against unlawful attacks, but in this case the attacking soldier’s original assault is lawful under LOAC as long as the anticipated collateral damage is not disproportionate to the value of the military objective. If the envisioned collateral damage is disproportionate, then the attack is unlawful.  So, in that case, the right of the civilian to exercise self-defense would depend entirely on the civilian making the correct assessment of not only the collateral damage, but also the anticipated collateral damage viewed — not from his perspective – but rather from the perspective of the person attacking him!  A greater conceptual riddle I cannot fathom.  This would be a nightmare for a criminal court, international or domestic, to adjudicate.

The deeper issue imbedded in Henderson and Cavanaugh’s research is the application of the privilege of combatancy to non-international armed conflicts.  The standard textbook answer is that the privilege is inapplicable to NIAC because the very concept of “combatant” is part of the legal architecture of IAC.  Under this view, a NIAC can only have government forces and rebels – never combatants per se.

To my mind, this statement is often reflexively repeated in the literature without due consideration for whether it is always and universally true.  Few individuals have challenged it, though Henderson himself is one of the few to have seriously studied the issue, writing articles suggesting that government forces in NIACs are privileged belligerents and that prosecuting them domestically would violate the laws of war.  Henderson is to be commended for tackling an understudied but vital topic.

However, I might go even further than Henderson and Cavanagh, whose view is largely asymmetrical because it entails that only government forces (but never rebels) are entitled to the privilege.  In some circumstances I think we are too hasty to conclude that rebels can never be privileged (though I will subsequently qualify this statement below).

The standard argument for why rebels can’t be privileged is the last clause of Common Article 3 of the Geneva Conventions, which states that the protections therein do not alter the legal status of the parties.  Most people take from this statement the assumption that rebels, though protected from torture and other forms of ill treatment, remain unprivileged.  But Common Article 3 merely preserves the status quo, so the real question is the status of rebels prior to Geneva.

In the 19th Century, the classification scheme between conflicts was not as simplistic and dualist as today’s binary classification between IAC and NIAC.  The great publicists believed in a plethora of different type of internal conflicts, including disturbances, rebellions, and insurrections.  At some point such conflicts could ripen into full-blown civil wars that became the functional equivalent of international armed conflicts and triggered much of the machinery of international law.  It was therefore not true that civil wars were solely the province of sovereign prerogative; they could be matters of international concern.

Of course, this doesn’t mean that rebel leaders couldn’t be prosecuted for treason if the rebels were vanquished by the government.  They certainly could, because they violated an oath of loyalty to their own government when they decided to rebel. But as long as the rebels met the functional criteria for lawful belligerency (carrying arms openly and following the customs of war) wholesale prosecutions of regular rebel soldiers for domestic murder was disfavored – something that didn’t even happen during the U.S. Civil War.

The important point here is to keep separate the crime of treason from the crime of murder – which are based on totally different foundations. Treason is a betrayal of one’s duty of loyalty.  The privilege of combatancy negates the unlawfulness of the killing that would otherwise be murder, but it does nothing to negate the betrayal that stands at the core of treason.  And the same thing is true of IAC as well. If an American citizen joins the Russian Army and fights as a uniformed solider against the United States, he could be prosecuted for treason but the privilege of combatancy would prevent the United States from prosecuting him for murder committed on the battlefield.

All of this suggests that the modern classification scheme has lost the richness of these older conceptual categories. And we might have been too hasty in concluding that rebels can never enjoy the privilege of combatancy.  In some situations it may be logically coherent to extend the privilege of combatancy to them during some non-international armed conflicts – and at the same time prosecute them for treason and other loyalty-based offenses if they have violated a pertinent duty or oath to their sovereign.

23 Responses

  1. Because this country owes its existence in part to men like Lafayette who helped George Washington train his men, the US broke with British common law and allows its citizens to join foreign armies not at war with the US. During the two years of WWII when the US was neutral, thousands of US citizens joined the armies on both sides. We celebrate the men who made the right choice: three squadrons of fighter pilots who fought for the RAF during the Battle of Britain, the AVG (Flying Tigers) who fought the Japanese for China. We don’t say much about the thousands of German American and Italian American boys who ended up on the wrong side of the war. None of them were prosecuted for treason. If you join a foreign army when the US is at peace, and then you are stuck in that army if it goes to war with the US, then you did nothing illegal and what happened is not under your control. You don’t get out of an army until they let you out, and while you are in an army you have to follow orders. However, if you join an army already at war with the US, then you commit treason.

    It would be peculiar if the US decided not to charge Confederates who rebelled against the government with treason while prosecuting some guy who had the bad luck to join the Finnish army fighting Russians in 1940 (Finland and Thailand are the two least commonly remembered Axis countries in WWII).

  2. Jens,

    Thank you for the thoughtful post and your comments. Needless to say, we agree with almost all of what you have written.

    To the limited extent that we disagree, we would still argue that the non-government forces in a NIAC lack the equivalent of the combatant’s privilege. Your point about the last sentence of GCIII is well taken, and we recognise that in older times the legal landscape was a more complicated and diverse place. We also point to article 6(5) of APII.

    Perhaps the best way to resolve the question as to what is the current state of the law would be to see if States have prosecuted members of rebel or insurgent forces who, while complying with the principle of distinction (eg, by wearing affixed distinctive sign visible at a distance) and employing lawful weapons, killed members of the government armed forces. It would also be informative if other States were aware of such prosecutions and did not object or otherwise register their disagreement.

    Are any of the readers aware of such prosecutions?

  3. In any context in which a civilian has the opportunity to take up arms against a military attacker, the same person has an opportunity to fall to his knees and put his hands behind his head. Self defense does not necessarily apply when the attacker is bound by the principle of discrimination. In order for discrimination to work and to be generally honored, civilians are obligated to take no part in combat even when they feel themselves threatened. Therefore, because of the option of surrender and for the protection of discrimination, neither of which apply when one civilian is attacked by another civilian, “self defense” and other civilian concepts may not excuse direct participation in hostilities by an unprivileged combatant.

  4. Howard,

    What about Jehns’s scenario — Soldier S is attacking Enemy E with heavy weapons, which exposes Civilian C to injury or death. In what circumstances, if any, can C ‘defend’ themsleves against S?

  5. Howard, not sure I understand the last point. Let me restate the basic point here: although a citizen does not have the right to participate in hostilities, he retains the right to exercise self-defense against an unlawful attack against him. The unlawful attack against him can be unlawful because the attacker is unprivileged, or because the attack violates another prohibition of jus in bello.

  6. If the attacker is unprivileged then this is a standard case of attempted murder and self defense. Otherwise, war is a matter for the military on both sides. Civilians may not be attacked, but to protect that principle civilians may not participate. However, if an attacking military commits a war crime, particularly the murder of civilians, then self defense by other civilians is a plausible claim. For other claims that “the attack violates another prohibition” that is for the military courts to sort out later, just as citizens cannot claim that they had a right to shoot SWAT officers due to some alleged defect in the warrant.

    Given the recent nonsense alleged by lawyers for Omar Khadr, it is important to remember that non privileged belligerency is an charge that is only made by the enemy and tried in enemy military courts. You never charge your own people with killing the enemy. No Frenchman in the resistance was ever charged with murder for killing German soldiers. So the only time this ends up in civilian courts is after rebellion where issues of jus in bello don’t typically work against the lawful government.

  7. Part of the confusion, evident even partly here, is that domestic self-defense law that will be used in a domestic trial depends on the domestic law of the relevant domestic legal process whereas combatant immunity depends on international law, CIL in particular. The significant issue is how international law is, can be, incorporated into a particular domestic legal process re: a dometic trial for murder. In the United States, the Supreme Court, especially in dictum in Ex parte Quirin, has used direct incorporation(i.e., international law directly re: status, rights, duties, competencies — as the S.Ct. stated famously in Ex parte Quirin [even though the S.Ct. was applying legilation that incorporated all of the laws of war as offenses against the laws of the United States by reference, and for other reasons] [see ]
    I take issue with any notion that a member of the regular armed forces of a party to an international armed conflict {Yes, David, including those with a “belligerent” status during the U.S. Civil War — and, yes, David today it would be an IAC0 has to have a distinctive insignia recognizable at a distance 9see ). All that was required under the Lieber Code and the 1907 HC No. IV (see, e.g., id.) was membership in the armed forces.
    Howard: and the U.S. prosecuted members of the army of the United States as well as members of the army of the “belligerent” CSA for violations of the laws of war. A presidential amnesty for treason was issued, but it did not cover war crimes. Moreover, case law recognized that U.S. citizenship had not been lost or changed re: members of the army of the CSA — which is interesting re: another matter — whether U.S. citizenship requires a special immunity from drone targeting during war [no] or targeting during combat during WWII or freedom from being detained as a pow, etc. [no] or required some special notice before being targeted [no] — see [an earlier draft is posted and the editors do not want me to post the present working draft, which has more U.S. cases]

  8. p.s. the S.Ct. case involved the state of Texas (no loss of citizenship, remained subject to U.S. laws).

  9. Jordan, why is an insignia not required?

  10. First, I apologize Jens (for using “David” — sounds like Mary Ellen?).
    The Yale J. article demonstrates why, but the 1863 Lieber Code recognized that so soon as a person takes the oath (membership), that person has the combatant’s privilege. HC No. IV, annex, art. 1 contains the famous “or” — thus demonstrating that for related pow status membership is the sole criterion for members of the regular armed forces or “armies.” GPW, art. 4(A)(2) criteria are expressly relevant to a very limited set of folk and the insignia element is not relevant re: 4(A)(1) or (3). State practice confirms that members of the regular armed forces that were camouflage are still combatants and are still entitled to pow status under 4(a)(1). The U.S. is not bound by GPI.

  11. It seems to me that the four requirements for belligerency apply to regular armed forces as well. The reason that they are explicitly listed for militias and volunteer groups is that these groups become the functional equivalent of regular armed forces when they meet this criteria. Everyone assumes that regular armed forces meet these criteria which is why these criteria are used to determine which irregular forces are legitimate combatants.

  12. of course, I meant wear camouflage.
    The problem is that of all six categories in 4(A) of GPW, only 4(a)(2) contains the list — and one must interpret the treaty thusly. “Everyone assumes” is not easy to prove, but it is actually well-recognized that membership in the regular armed forces is the sole criterion. JAG Operational Handbook, etc.–please see the Yale J. article.
    State practice appears to confirm this re: use of camouflage (to look like the snow, the forest, etc.).

  13. Jordan – Yes the US is obligated to prosecute its own people for “war crimes”, particularly serious violations of international law. However, there are less serious violations of the laws of war that we prosecute selectively against the enemy while the enemy only prosecutes our people. For example, “crossing lines without uniform” (being a military spy) is something the US prosecutes the enemy for doing (Quirin) but it is also something we do (OSS, SOE)and when we do it we do not prosecute our people for having violated the laws of war against the enemy. Our first great American hero, Captain Nathan Hale, violated this law of war but we erect statues and name Grade Schools after him, we do not call him a “war criminal” because his offense against the laws of war was not a real “war crime”.

    It is in this context that we prosecute unprivileged belligerents who kill American soldiers (Omar Khadr) but we do not prosecute unprivileged belligerents who are US agents killing enemy soldiers (CIA drone pilots but also 1980’s Afghan civilians with CIA Stinger missiles shooting down Soviet helicopters). In the same way, the US prosecuted KGB agents for espionage when we caught them while at the same time tasking the CIA to conduct espionage against the Soviets.

    The non-crime violations of the laws of war also have peculiar rules. Someone who crosses lines without uniform cannot be punished if he successfully returns back through the lines to his own forces. You can punish a POW for attempting to escape, but if he is successful and gets back to his own lines then you can no longer punish him for the successful escape. So obviously in neither case do we punish someone who on our behalf successfully committed what the enemy is free to consider a violation of the laws of war after the enemy is no longer allowed to prosecute the offense due to a different rule in the same laws of war.

  14. Howard: I believe that Ex Parte Quirin involved enemy combatants who were in a process of attack (sabotage) while out of uniform, which is the war crime. It is not a war crime to kill U.S. soldiers while one is unprivileged — one merely does not have the privilege to do so and is subject to being prosecuted under a relevant domestic law for murder. I realize that the GTMO milt. comms. apparently do not understand this.
    Regarding the CIA fliers of drones, May Ellen and I, among others, have pointed out that they would not be combatants under the laws of war (unless they were also members of the regular armed forces of the U.S.), but I had written that practice and opinion seems to provide them with sn immunity under the law of self-defense for lawful targetings under the law of self-defense.

  15. p.s., I mean kill U.S. soldiers in combat — the killing of any detainee of any status can be a war crime. But this brings us back to the postings: the killing of a detainee who acquires a weapon and is about to shoot?

  16. “subject to being prosecuted under a relevant domestic law for murder” I recognize that this is the usual formula, but in the recent conflict you run up against the fact that a civilian in Afghanistan killing a US soldier (Omar Khadr killing Sargent Speer) committed no crime in Afghan law as it stood at that time. In Afghanistan, the municipal law held that anyone was free to kill Americans (military or civilian) any time anywhere. So courts cannot enforce “domestic law” in a country like that. They have to assert some generally CIL common law prohibition against murder and prosecute under that charge. That the crime can be charged and prosecuted seems obvious, but whether in this case the right charge was brought is plausibly in dispute.

    The charge in Quirin was crossing lines without uniform. It is true that if they had committed an act of sabotage then that would be an independent crime and one that doesn’t disappear after the spy returns to his own lines. However, intent to commit sabotage without the actual sabotage is not an additional charge.

  17. Jordan, consider the following hypothetical: The international state of Hypothetica creates a Department of Defense with a branch called the Armed Forces of Hypothetica. The soldiers in this armed force do not wear a fixed insignia, do not have a command structure, do not carry their arms openly, and do not follow the laws and customs of warfare. Would they be entitled to POW status and the privilege of belligerency?

  18. “In the criminal law we usually view the right of self-defense as only applicable against unlawful attacks, but in this case the attacking soldier’s original assault is lawful under LOAC as long as the anticipated collateral damage is not disproportionate to the value of the military objective. If the envisioned collateral damage is disproportionate, then the attack is unlawful. So, in that case, the right of the civilian to exercise self-defense would depend entirely on the civilian making the correct assessment of not only the collateral damage, but also the anticipated collateral damage viewed — not from his perspective – but rather from the perspective of the person attacking him! A greater conceptual riddle I cannot fathom. This would be a nightmare for a criminal court, international or domestic, to adjudicate.”

    I wanted to tease out reflection on the ordinary civilian in this space.

    Hypo 1, the combatant makes a lawful attack on a military objective and in doing that their fire hits a civilian who would be properly classified as collateral damage (i.e. not a war crime by the combatant).

    Hypo 2, the combatant makes an unlawful attack with fire that hits a civilian.

    In either case, from the civilian’s perspective, they are being fired upon by the combatant.

    Assume also there is no possibility of fleeing for the civilian. Also there is no possibility of waving some flag of surrender due to the uncertainty as to whether the combatant will accept the surrender or would even acknowledge the indicia of willingness to surrender. Assume also that the civilian has not in any manner taken part in these hostilities – a true innocent bystander in an area of armed conflict.

    Finally, like in the United States now – assume civilians carry weapons including guns regularly as part of going about their business.

    Does the civilian have a right of self-defense vis a vis the combatant or not?

    It would seem to me perverse if we were to say that the laws of war mean that a civilian does not have a right of self-defense to shoot back at that soldier whether the soldier is making a lawful or unlawful attack. It is too much for the law to ask that civilian to proceed as a lamb to the slaughter.

    If we think in terms of international human rights law, it would also seem perverse to say that civilian would have no recourse.

    Does that mean that the civilian becomes DPH because they fire back? They have begun participation in some hostilities but are they participating in “THE” hostilities. Maybe yes and because of that they may lose the immunity of the innocent bystander and become targetable by the combatant.

    If that civilian is captured, could they be charged with a war crime “murder inconsistent with the laws of war”. Well the power that created the combatant could certainly take the view that no one should fire back even when fired upon, but as a matter of law should that be the only answer possible? It would seem to me that the civilian could argue necessity or duress to the charge of a war crime – they were not party to the conflict but were brought into it by the action of the combatant which – whether lawful or not put the civilian in the place of necessity or duress leading to the civilian firing back in self-defense.

    The civilian could also be charged with domestic law murder where they would assert a right of self-defense of self or another (imagine their family was in the line of fire also).

    What result at the end of the day? I can see many different results from guilt to acquittal.

    The thing for me is that the civilian should have a right of self-defense as it has the salutary effect of raising – in a very practical manner – the risk for the combatant in causing collateral damage or doing an unlawful attack. This is totally consistent with the relevant principles that are to guide the use of force by the combatant.

    Obviously, I am just thinking these things through and do not have a definite answer, but I wanted to fathom them a bit more here as I think about the civilians caught in a cross-fire and trying to just survive – a fairly basic human instinct that I can not imagine the law does not recognize in an IAC or a NIAC setting.

    But, of course, I may just be wrong.


  19. “So, in that case, the right of the civilian to exercise self-defense would depend entirely on the civilian making the correct assessment of not only the collateral damage, but also the anticipated collateral damage viewed — not from his perspective – but rather from the perspective of the person attacking him! A greater conceptual riddle I cannot fathom. This would be a nightmare for a criminal court, international or domestic, to adjudicate.”

    Combatant who is lawfully or unlawfully shooting at the innocent bystander who is a civilian, 1) who is reasonable in not trusting that kneeling and putting up his arms will lead to him or his family not being shot, 2) is unable to escape and 3) who carries a gun like everyone does in that place or country.

    That to me is a more accurate image of the civilian’s dilemma. Does the law of war allow that civilian to shoot back at the combatant shooting at them?

    If no, what is the civilian supposed to do – just let themselves be killed? If not, what is the basis for them being able to shoot back – where is that derived from.

    I have read the other posts here and have not found a satisfactory answer for the situation of the civilian innocent bystander constrained to shoot or die. Is there something in levee en masse when one is alone but acting spontaneously in this setting not to participate in the hostilities but to save one’s own life? self-defense, necessity, or duress based excuses for the killing?

  20. Jens – This depends on who you talk to. Historically the four part test applies to militia units operating independently from the main army. When the Geneva Conventions were being negotiated the US proposed extending it to all armed forces, but the Russians pointed out that this would be at variance from the Hague Agreements of 1907 and the original design of the Geneva Convention was to always agree with Hague. So the final wording kept the status quo. However, whether it was simply assumed that the main army would follow the rules or whether it did not have to follow the rules is a matter of controversy.

    This isn’t hypothetical. It was the Marxist uniformed Afghan Army that invited the Soviets into the country, so in the 1980’s everyone in uniform was a bad guy. Therefore, in the 90’s the Taliban army did not wear uniforms and following Afghan custom they did not have ranks or a unified chain of command. This led Yoo and Bellinger to separately author opinions that Afghanistan had no army but instead had only a massed tribal militia that did not meet the four part test. Therefore the enemy was not entitled to POW protection and was subject to enhanced interrogation.

    However, if the Afghan army was an army with attached militia and the four part test applies only to detached militia, then prisoners should have been afforded POW status and any enhanced interrogation was illegal. So this is a very important question even in the cases still active today.

  21. Howard, thank you for this very informative reply.

  22. Ben: the laws of war leave the civililan self-defender unprivileged and subject to relevant domestic law or other relevant international law. Room for change? Would change be prefereable?
    Howard: in other posts here, I found it interesting the the European Convention on H.R. seems to have a built-in right of self-defense and defense of others. And Ben, this could be relevant, of coourse, re: parties to that treaty. I suspect that it would be useful to explore patterns of practice and opinio juris re: whether there is a customary human right of personal self-deffense and/or defense of others. Documenting widespread domeistic law does not help re: the final issue: is that practice operative along with a generally shared expectation that it is part of international law (opinio juris)? Another part of a future study?
    Jens: if they are members of the regular armed forces of that state, yes.

  23. Under what legal category in the theory of just war were the Apache, Shoshoni, Herero and Xosa?

Trackbacks and Pingbacks

  1. There are no trackbacks or pingbacks associated with this post at this time.