Human Shields in International Humanitarian Law and in Domestic Criminal Law
by Michael Dorf
As I post these thoughts, there is no ceasefire in place to stop the immediate bloodshed in Gaza and Israel, much less any apparent progress towards resolving the larger conflict. But, as with my most recent post inspired by the latest outbreak of violence, I am going to address a general issue it raises, rather than assess the current situation or the broader conflict. I realize that in treating a very real tragedy as merely the insipiration for an intellectual discourse, I risk coming across as detached or unfeeling, so let me assure readers that I take very seriously both Israel's security concerns and the devastation in Gaza. As with my last post on the current Israel/Hamas conflict, I want to discourage comments on topics other than the one I address here, and to announce in advance that I won't respond to any such off-topic comments.
With that warning/disclaimer out of the way, I come to the question that concerns me: Can we learn anything about the human shields question in international humanitarian law (IHL) by looking at domestic criminal law?
The issue is inspired by a question that I bracketed in my last post on civilian casualties: If Force A uses human shields, does that affect the determination of whether Force B's use of force against A is proportionate (and thus legal)? Israel says that Hamas deliberately embeds itself in the civilian population, storing and launching rockets from, and digging tunnels under schools, hospitals, and mosques, as well as preventing some Gaza civilians from heeding Israeli warnings to evacuate targeted buildings--and that therefore, the proportion of civilian casualties from Israel's attacks on Hamas should be permissible even if that proportion is higher than the threshold that would lead to the conclusion that Israel would be violating the proportionality norm absent the deliberate use of human shields.
One could contest various premises here. For one thing, Hamas is not exactly the government of a country, and so the application of IHL to it is open to question. Nor is it entirely clear that IHL, as opposed to the stricter norms applicable to occupying forces, applies to Israel with respect to Gaza. Although Israel dismantled all of its Gaza settlements in 2005, Gaza remains subject to considerable Israeli (and also Egyptian) control. In addition, Hamas and others argue that the population density of Gaza, rather than its deliberate policies, are mostly responsible for the high civilian casualty rates.
These and other issues would need to be addressed in any assessment of war crimes liability for Hamas and/or Israel, but I'm going to put them aside here to focus on a general issue that arises in many asymmetrical conflicts: Does the measure of proportionality in determining the lawfulness of attacks that foreseeably lead to collateral civilian casualties change when the enemy uses civilians as human shields? (There is some disagreement about what exactly it means for a force to use human shields, but I use the term loosely to refer to any actions deliberately undertaken to ensure that attacks by the enemy on military targets will likely result in substantial civilian casualties.)
Since writing my last post, I did a bit of research and came to the conclusion that IHL is not fully determinate on the question, but is probably best read as not changing its definition of proportionality based on the use of human shields by the enemy. The basic idea appears to be that war crimes by one side don't justify war crimes by the other.
That logic has been questioned. For example, a Student Note in the 2012 Vanderbilt Journal of Transnational Law argues that IHL should give the attacking side the right to disregard human shields entirely, on the ground that this approach, if followed consistently, would reduce the incentive for the enemy to use human shields in the first place. For what it's worth, I think that's probably wrong even on its own terms. Even if foreseeably killing unlimited numbers of human shields were lawful under IHL, many countries would refrain from doing so out of moral and/or public relations considersations, and so there would remain an incentive for the other side in a conflict to use human shields.
A less extreme reform proposal comes from Amnon Rubinstein and Yaniv Roznai in the 2011 Stanford Law and Policy Review. They argue first that IHL ought to be clearer that forces deploying human shields commit grievous war crimes. Next, they urge distinguishing between voluntary human shields and involuntary ones; people who volunteer to serve as human shields should not count as civilians for purposes of calculating proportionality (as civilians participating in hostilities ordinarily do not so count during the period of their participation); but given the difficulty of determining whether particular individuals are voluntary or involuntary shields, uncertainty should be resolved with a presumption in favor of treating human shields as involuntary, and thus protected. Finally, Rubinstein and Roznai would retain the proportionality requirement as against an enemy employing involuntary human shields, but they would relax it somewhat. Thus (and this is my account of their proposal), if for some military operation, the proportionality limit on the number and severity of civilian casualties would be X in the absence of human shields, it would be X+Y if the enemy is using involuntary human shields.
I think there is much to be said for the Rubinstein/Roznai proposal but also much to be said for the two-wrongs-don't-make-a-right logic of the current IHL norm (as I perceive it). I do not start out with a strong view about which approach is to be preferred in IHL, but perhaps we can get some guidance from the parallel question in domestic criminal law.
Hostage-taking is a close cousin of the use of human shields in domestic criminal law, but its usual logic is importantly different. If bank robbers hold customers and bank employees hostage in order to facilitate their own escape, they are using the hostages as shields, but the robbers usually aren't threatening violence to anyone else, so it is difficult to imagine a scenario in which one might think that intentionally killing the bank robbers while foreseeably but regrettably killing the hostages makes any sense.
It is difficult to come up with a realistic scenario in domestic crime that closely parallels the IHL case. (For an unrealistic scenario, I recommend the boat scene from the Batman film The Dark Knight). But realistic or not, here is what I have in mind: Bad Guy, with the intent to kill Good Guy, has a gun pointed at Good Guy, who also has a gun, but Bad Guy is holding Innocent Shield in front of him, so that the only way that Good Guy can shoot Bad Guy before Bad Guy shoots Good Guy (who cannot safely retreat) is by shooting in a way that will foreseeably (albeit regrettably) harm or kill not just Bad Guy but also Innocent Shield. Is Good Guy permitted to shoot?
Notice that in my hypothetical example, Good Guy wants to use deadly force to protect himself. I am not asking whether a police officer would be justified in shooting. I think the answer to that question would generally be no, even if Good Guy would be permitted to shoot: A police officer would have no reason to prefer the life of Good Guy to Innocent Shield--although a police officer might have reason to think that there's at least a chance that he could hit Bad Guy without harming Innocent Shield, and so conclude that there will be a lower chance of the loss of innocent life if he shoots than if he doesn't. But let's put that issue aside. Although one could analogize the use of force by countries in war to the use of force by police, more commonly national self-defense is analogized to individual self-defense.
So, can Good Guy use deadly force against Bad Guy, knowing that there is a substantial probability that in doing so he will also kill Innocent Shield? Put differently, suppose Good Guy does use deadly force in this way, resulting in the death of both Bad Guy and Innocent Shield. Will he face criminal liability?
Let's look at the answer under the New York Penal Law (which pretty closely follows the Model Penal Code; the answer may differ in other states, but not that much, I suspect). Prima facie, Good Guy has committed two second-degree murders because he acted intentionally with respect to Bad Guy and at least extremely recklessly with respect to Innocent Shield. (Alternatively, Good Guy might be on the hook for Innocent Shield under a "transferred intent" theory.) Under NY Penal Code Sec. 35.15, Good Guy can make out a successful defense of justification (i.e., self-defense) with respect to Bad Guy, but not with respect to Innocent Victim, because self-defense only applies against the attacker himself. Is there some other defense available?
Duress is arguably a possibility. In the standard duress scenario, X uses duress to induce Y to commit an offense against Z. (E.g., "mug that old lady or I'll break your knees"), but we can imagine that the threat that Bad Guy poses to Good Guy would be the duress that leads Good Guy to kill Innocent Shield. (Joshua Dressler proposes that Model Penal Code duress would be an appropriate defense in the related case where a battered woman kills her batterer while he is sleeping.) Maybe this fits the literal language of the duress defense: "the defendant engaged in the proscribed conduct because he was coerced to do so by the use or threatened imminent use of unlawful physical force upon him or a third person, which force or threatened force a person of reasonable firmness in his situation would have been unable to resist." The answer would depend on what "coerced to do so" means. Certainly Bad Guy is not trying to get Good Guy to kill Innocent Shield; Bad Guy is using Innocent Shield to prevent Good Guy from shooting at Bad Guy; so a duress defense only works if we imagine that it includes unintended coercion.
Another possibility would be the defense of necessity. NY statutory law does not expressly recognize the defense of necessity, but it is encompassed within the general provision recognizing justification defenses. However, that provision by its terms excludes those defenses relating to the use of physical force, so, in New York, at least, necessity is unavailable. This reasoning also suggests that a duress defense would be unavailable. The legislature specifically considered the use of force in self-defense and concluded in Sec. 35.15 that it should only be permitted against the attacker, not against innocent third parties. Given the canon that the specific supersedes the general, it would be incongruous to say that nonetheless Good Guy can make out a defense of duress.
Thus, while recognizing that I'm not an expert in criminal law, I conclude that NY law--and the law of other states to the extent that it is similar--would not provide for a valid defense for Good Guy's foreseeable killing of Innocent Shield in my hypothetical example. Is that result normatively justified?
I think the answer is probably yes, for two reasons. First, on utilitarian grounds, it's worse to kill two people than one (even if one of the two is a criminal).
Second (and more importantly for me), on deontological grounds, it could be thought impermissible to deliberately act (as opposed to failing to act) in a way that causes severe harm to another, even if doing so avoids an equally severe harm. That is why many deontologists think it impermissible to switch the trolley from one track to another, even if they thereby avert a greater harm. Whether one agrees with that conclusion, even as a deontologist, depends on how seriously one takes the doctrine of double effect. Here, the killing of Innocent Shield is not the intended consequence of the act, just the regrettable side effect. Good Guy, in this view, is not using Innocent Shield in the way that the "fat man" is used in a variant on the trolley problem that nearly all deontologists find objectionable. Therefore, under the double effect doctrine, Good Guy acts permissibly if the incidental harm to Innocent Shield is proportionate to the harm averted. Double effect, in moral theory, as in IHL, requires both that the harm to innocents be collateral and proportionate. If you find this brief discussion of "trolleyology" too brief to get a handle on it, you might want to check out my earlier post on the subject.
What's notable here is that New York's criminal law does not appear to permit the foreseeable but unintended killing of innocents even when that killing is proportional. So if instead of threatening one person, Bad Guy were threatening two, or five, or a hundred, so long as the only way for one of those two, or five, or a hundred to disable Bad Guy is to shoot through Innocent Shield, as I read the criminal law, that would still be impermissible, even though it would be permissible for some deontologists under the double effect doctrine and permissible for utilitarians on the ground that it would save more lives. Yet domestic criminal law appears to categorically forbid the collateral killing of human shields. With respect to collateral harm to innocents, domestic criminal law is much more demanding of those who would use force in personal self-defense than IHL is of nations using force in war.
To be sure, in the domestic case Good Guy (or the two, five, or a hundred good guys in the variants) could still go ahead and shoot, hoping that a prosecutor will exercise her discretion not to bring charges (or to reduce the charges), that a grand jury will refuse to indict, or that a jury will nullify. But then, there are also mechanisms by which IHL is under-enforced. For many people who have committed war crimes, the de facto punishment amounts to no more than a foreign travel ban, because they remain powerful in their home countries.
If one thinks that the domestic law treatment of human shields is appropriate, then one will be tempted to say that the reform proposals to loosen the proportionality requirement in the face of human shields should be rejected. IHL already licenses more force against innocents than domestic criminal law, after all. But even if one thinks that the domestic law treatment of human shields (as I have described it) is too tough on the person placed in the difficult situation of needing to (regrettably and collaterally) take an innocent life to save his own, then at most one will end up concluding that the existing IHL norm allowing proportionate collateral harm to innocents is appropriate, but one will need a further argument for loosening that norm based on the enemy's use of involuntary human shields. Such a further argument would have to explain how and why nations should be given greater leeway to inflict collateral harm against human shields than individuals should be given. I suppose such an argument could be made, but I haven't seen it.
As I post these thoughts, there is no ceasefire in place to stop the immediate bloodshed in Gaza and Israel, much less any apparent progress towards resolving the larger conflict. But, as with my most recent post inspired by the latest outbreak of violence, I am going to address a general issue it raises, rather than assess the current situation or the broader conflict. I realize that in treating a very real tragedy as merely the insipiration for an intellectual discourse, I risk coming across as detached or unfeeling, so let me assure readers that I take very seriously both Israel's security concerns and the devastation in Gaza. As with my last post on the current Israel/Hamas conflict, I want to discourage comments on topics other than the one I address here, and to announce in advance that I won't respond to any such off-topic comments.
With that warning/disclaimer out of the way, I come to the question that concerns me: Can we learn anything about the human shields question in international humanitarian law (IHL) by looking at domestic criminal law?
The issue is inspired by a question that I bracketed in my last post on civilian casualties: If Force A uses human shields, does that affect the determination of whether Force B's use of force against A is proportionate (and thus legal)? Israel says that Hamas deliberately embeds itself in the civilian population, storing and launching rockets from, and digging tunnels under schools, hospitals, and mosques, as well as preventing some Gaza civilians from heeding Israeli warnings to evacuate targeted buildings--and that therefore, the proportion of civilian casualties from Israel's attacks on Hamas should be permissible even if that proportion is higher than the threshold that would lead to the conclusion that Israel would be violating the proportionality norm absent the deliberate use of human shields.
One could contest various premises here. For one thing, Hamas is not exactly the government of a country, and so the application of IHL to it is open to question. Nor is it entirely clear that IHL, as opposed to the stricter norms applicable to occupying forces, applies to Israel with respect to Gaza. Although Israel dismantled all of its Gaza settlements in 2005, Gaza remains subject to considerable Israeli (and also Egyptian) control. In addition, Hamas and others argue that the population density of Gaza, rather than its deliberate policies, are mostly responsible for the high civilian casualty rates.
These and other issues would need to be addressed in any assessment of war crimes liability for Hamas and/or Israel, but I'm going to put them aside here to focus on a general issue that arises in many asymmetrical conflicts: Does the measure of proportionality in determining the lawfulness of attacks that foreseeably lead to collateral civilian casualties change when the enemy uses civilians as human shields? (There is some disagreement about what exactly it means for a force to use human shields, but I use the term loosely to refer to any actions deliberately undertaken to ensure that attacks by the enemy on military targets will likely result in substantial civilian casualties.)
Since writing my last post, I did a bit of research and came to the conclusion that IHL is not fully determinate on the question, but is probably best read as not changing its definition of proportionality based on the use of human shields by the enemy. The basic idea appears to be that war crimes by one side don't justify war crimes by the other.
That logic has been questioned. For example, a Student Note in the 2012 Vanderbilt Journal of Transnational Law argues that IHL should give the attacking side the right to disregard human shields entirely, on the ground that this approach, if followed consistently, would reduce the incentive for the enemy to use human shields in the first place. For what it's worth, I think that's probably wrong even on its own terms. Even if foreseeably killing unlimited numbers of human shields were lawful under IHL, many countries would refrain from doing so out of moral and/or public relations considersations, and so there would remain an incentive for the other side in a conflict to use human shields.
A less extreme reform proposal comes from Amnon Rubinstein and Yaniv Roznai in the 2011 Stanford Law and Policy Review. They argue first that IHL ought to be clearer that forces deploying human shields commit grievous war crimes. Next, they urge distinguishing between voluntary human shields and involuntary ones; people who volunteer to serve as human shields should not count as civilians for purposes of calculating proportionality (as civilians participating in hostilities ordinarily do not so count during the period of their participation); but given the difficulty of determining whether particular individuals are voluntary or involuntary shields, uncertainty should be resolved with a presumption in favor of treating human shields as involuntary, and thus protected. Finally, Rubinstein and Roznai would retain the proportionality requirement as against an enemy employing involuntary human shields, but they would relax it somewhat. Thus (and this is my account of their proposal), if for some military operation, the proportionality limit on the number and severity of civilian casualties would be X in the absence of human shields, it would be X+Y if the enemy is using involuntary human shields.
I think there is much to be said for the Rubinstein/Roznai proposal but also much to be said for the two-wrongs-don't-make-a-right logic of the current IHL norm (as I perceive it). I do not start out with a strong view about which approach is to be preferred in IHL, but perhaps we can get some guidance from the parallel question in domestic criminal law.
Hostage-taking is a close cousin of the use of human shields in domestic criminal law, but its usual logic is importantly different. If bank robbers hold customers and bank employees hostage in order to facilitate their own escape, they are using the hostages as shields, but the robbers usually aren't threatening violence to anyone else, so it is difficult to imagine a scenario in which one might think that intentionally killing the bank robbers while foreseeably but regrettably killing the hostages makes any sense.
It is difficult to come up with a realistic scenario in domestic crime that closely parallels the IHL case. (For an unrealistic scenario, I recommend the boat scene from the Batman film The Dark Knight). But realistic or not, here is what I have in mind: Bad Guy, with the intent to kill Good Guy, has a gun pointed at Good Guy, who also has a gun, but Bad Guy is holding Innocent Shield in front of him, so that the only way that Good Guy can shoot Bad Guy before Bad Guy shoots Good Guy (who cannot safely retreat) is by shooting in a way that will foreseeably (albeit regrettably) harm or kill not just Bad Guy but also Innocent Shield. Is Good Guy permitted to shoot?
Notice that in my hypothetical example, Good Guy wants to use deadly force to protect himself. I am not asking whether a police officer would be justified in shooting. I think the answer to that question would generally be no, even if Good Guy would be permitted to shoot: A police officer would have no reason to prefer the life of Good Guy to Innocent Shield--although a police officer might have reason to think that there's at least a chance that he could hit Bad Guy without harming Innocent Shield, and so conclude that there will be a lower chance of the loss of innocent life if he shoots than if he doesn't. But let's put that issue aside. Although one could analogize the use of force by countries in war to the use of force by police, more commonly national self-defense is analogized to individual self-defense.
So, can Good Guy use deadly force against Bad Guy, knowing that there is a substantial probability that in doing so he will also kill Innocent Shield? Put differently, suppose Good Guy does use deadly force in this way, resulting in the death of both Bad Guy and Innocent Shield. Will he face criminal liability?
Let's look at the answer under the New York Penal Law (which pretty closely follows the Model Penal Code; the answer may differ in other states, but not that much, I suspect). Prima facie, Good Guy has committed two second-degree murders because he acted intentionally with respect to Bad Guy and at least extremely recklessly with respect to Innocent Shield. (Alternatively, Good Guy might be on the hook for Innocent Shield under a "transferred intent" theory.) Under NY Penal Code Sec. 35.15, Good Guy can make out a successful defense of justification (i.e., self-defense) with respect to Bad Guy, but not with respect to Innocent Victim, because self-defense only applies against the attacker himself. Is there some other defense available?
Duress is arguably a possibility. In the standard duress scenario, X uses duress to induce Y to commit an offense against Z. (E.g., "mug that old lady or I'll break your knees"), but we can imagine that the threat that Bad Guy poses to Good Guy would be the duress that leads Good Guy to kill Innocent Shield. (Joshua Dressler proposes that Model Penal Code duress would be an appropriate defense in the related case where a battered woman kills her batterer while he is sleeping.) Maybe this fits the literal language of the duress defense: "the defendant engaged in the proscribed conduct because he was coerced to do so by the use or threatened imminent use of unlawful physical force upon him or a third person, which force or threatened force a person of reasonable firmness in his situation would have been unable to resist." The answer would depend on what "coerced to do so" means. Certainly Bad Guy is not trying to get Good Guy to kill Innocent Shield; Bad Guy is using Innocent Shield to prevent Good Guy from shooting at Bad Guy; so a duress defense only works if we imagine that it includes unintended coercion.
Another possibility would be the defense of necessity. NY statutory law does not expressly recognize the defense of necessity, but it is encompassed within the general provision recognizing justification defenses. However, that provision by its terms excludes those defenses relating to the use of physical force, so, in New York, at least, necessity is unavailable. This reasoning also suggests that a duress defense would be unavailable. The legislature specifically considered the use of force in self-defense and concluded in Sec. 35.15 that it should only be permitted against the attacker, not against innocent third parties. Given the canon that the specific supersedes the general, it would be incongruous to say that nonetheless Good Guy can make out a defense of duress.
Thus, while recognizing that I'm not an expert in criminal law, I conclude that NY law--and the law of other states to the extent that it is similar--would not provide for a valid defense for Good Guy's foreseeable killing of Innocent Shield in my hypothetical example. Is that result normatively justified?
I think the answer is probably yes, for two reasons. First, on utilitarian grounds, it's worse to kill two people than one (even if one of the two is a criminal).
Second (and more importantly for me), on deontological grounds, it could be thought impermissible to deliberately act (as opposed to failing to act) in a way that causes severe harm to another, even if doing so avoids an equally severe harm. That is why many deontologists think it impermissible to switch the trolley from one track to another, even if they thereby avert a greater harm. Whether one agrees with that conclusion, even as a deontologist, depends on how seriously one takes the doctrine of double effect. Here, the killing of Innocent Shield is not the intended consequence of the act, just the regrettable side effect. Good Guy, in this view, is not using Innocent Shield in the way that the "fat man" is used in a variant on the trolley problem that nearly all deontologists find objectionable. Therefore, under the double effect doctrine, Good Guy acts permissibly if the incidental harm to Innocent Shield is proportionate to the harm averted. Double effect, in moral theory, as in IHL, requires both that the harm to innocents be collateral and proportionate. If you find this brief discussion of "trolleyology" too brief to get a handle on it, you might want to check out my earlier post on the subject.
What's notable here is that New York's criminal law does not appear to permit the foreseeable but unintended killing of innocents even when that killing is proportional. So if instead of threatening one person, Bad Guy were threatening two, or five, or a hundred, so long as the only way for one of those two, or five, or a hundred to disable Bad Guy is to shoot through Innocent Shield, as I read the criminal law, that would still be impermissible, even though it would be permissible for some deontologists under the double effect doctrine and permissible for utilitarians on the ground that it would save more lives. Yet domestic criminal law appears to categorically forbid the collateral killing of human shields. With respect to collateral harm to innocents, domestic criminal law is much more demanding of those who would use force in personal self-defense than IHL is of nations using force in war.
To be sure, in the domestic case Good Guy (or the two, five, or a hundred good guys in the variants) could still go ahead and shoot, hoping that a prosecutor will exercise her discretion not to bring charges (or to reduce the charges), that a grand jury will refuse to indict, or that a jury will nullify. But then, there are also mechanisms by which IHL is under-enforced. For many people who have committed war crimes, the de facto punishment amounts to no more than a foreign travel ban, because they remain powerful in their home countries.
If one thinks that the domestic law treatment of human shields is appropriate, then one will be tempted to say that the reform proposals to loosen the proportionality requirement in the face of human shields should be rejected. IHL already licenses more force against innocents than domestic criminal law, after all. But even if one thinks that the domestic law treatment of human shields (as I have described it) is too tough on the person placed in the difficult situation of needing to (regrettably and collaterally) take an innocent life to save his own, then at most one will end up concluding that the existing IHL norm allowing proportionate collateral harm to innocents is appropriate, but one will need a further argument for loosening that norm based on the enemy's use of involuntary human shields. Such a further argument would have to explain how and why nations should be given greater leeway to inflict collateral harm against human shields than individuals should be given. I suppose such an argument could be made, but I haven't seen it.