Follow Up to Bias Against Torture and Kind Right-Handers Imagine that one month after the 9/11 attacks the U.S. conducted a massive bombing campaign against terrorist training camps in Afghanistan. Further assume that (1) the U.S. knew Bin Laden was well hidden and so the attacks would have no significant chance of killing him, (2) the U.S. military had estimated that about 5,000 innocents would die because of the bombings and indeed around 5,000 innocents did die, and (3) the bombings were conducted mainly to deter future terrorist attacks against the U.S. Such a bombing campaign would have been widely supported in the U.S. Most Americans would assign the moral blame for the civilian deaths to Bin Laden.
I agree with Psychohistoran. Moreover, the second case could as well be "Let's kill five hundred Koreans" - there is no purpose in killing those who are not the enemy.
This really isn't a complicated problem. Civilian casualties as the result of an attack on a military target is considered acceptable. Civilian casualties caused for the sake of causing civilian casualties are not acceptable.
If the original hypothetical had some military purpose (say, we know 10 of these 510 family members are in fact top ranking al-Qaeda operatives, but we don't know which ones), then there would be a meaningful comparison between the two hypotheticals. As it stands, there does not appear to be much of a contradiction.
One can now, of course, begin constructing scenarios where the choice is between spending the entire world's GDP to defend yourself without losing an innocent life versus defending yourself for free at the cost of a single innocent life. I
That scenario will be difficult to construct. As soon as costs begin to get large they become equivalent to some number of innocent lives. Money represents resources, and as long as there are absolutely poor people, it's possible to spend small amounts of money in ways that prevent innocent deaths. Food and various public health measures cost money, and people die every day from their lack. Spending outrageous amounts of money on a war carries an opportunity cost in these lives, because even if we did none of these alternate things with the money explicitly, if it were simply not taxed, some small percentage of it would be donated to potentially lifesaving causes. The percentage spent and efficiency do not have to be very good to make a lot of lives saved out of what we've spent on the Iraq war, say. The percentages say that merely letting the taxpayers keep that money would have saved a whole bunch of lives. Actively using it with life-saving in mind would save a ridiculous number of lives.
In other words, the taboo against assassination exists entirely for the benefit of potential victims, who tend to be high-ranking and influential and famous and so able to perpetuate such a taboo.
Another version would be if the government could design a virus that would be carried asymptomatically by most people, but was designed to kill the host if they were genetically very similar to Osama bin Laden. Assuming such a virus could be made and was guaranteed not to kill anyone who wasn't at least a second cousin of Osama, what would be acceptable casualties?
In any case, I would hope that the modern US public disapproves of 'terror bombing', that is bombing designed primarily to cause human misery rather than to destroy the enemy's means of war. As far as I know the US hasn't used an explicit terror bombing strategy since WWII, though US bombing strategies in Vietnam are debatable on this score.
I always thought the taboo against assassination was because the thought process of military leaders (President/King/Emporer/Tsar/whatever) went something like this:
If we kill ten thousand of their soldiers, they will kill ten thousand of ours. That is an acceptable loss for our noble cause.
If we try to assassinate their leaders, they might try to asassinate me! That is not acceptable under any circumstances.
In other words, the taboo against assassination exists entirely for the benefit of potential victims, who tend to be high-ranking and influential and famous and so able to perpetuate such a taboo.
The best explanation is that we fundamentally tend to view directly intended acts differently from their forseen consequences (remember the Trolley Problem); so "conducting a bombing campaign with the knowledge that 5000 innocents will die" seems to fall under a different moral heading than does "assassinating 500 people".
To distinguish this from the hypothesis that it's about the specificity of the 500 family members, note that we would be just as appalled by an operation that rounded up 500 random Afghanis and killed them as a deterrent to Al Qaeda. The only thing that removes the moral outrage in the first case is our notion that intended acts are morally incommensurable with not-directly-intended consequences- a dangerous bias indeed, since those on the receiving end tend not to see the distinction so clearly.
This is the same exact mistake that Sam Harris makes in his book when he compares bombing casualties to torture. Perhaps they are morally comparable, but we are not judging the actions of individuals. We are judging the actions of governments. A government that can order torture and assassinations is a good deal more dangerous than a government that can order bombings. There are certain lines we must never let our governments cross.
My cynicism is such that I'm not sure that Bush would have been impeached. He'd have caused an almighty economic hoo-ha by murdering 500 members of one of the richest families of the country's leading oil producer, but I think he'd be able to present it as all the fault of the bad guys.
Killing 500 people with the surname Bin Laden would have made more sense than invading a country with no links to Al Qaeda - in fact, one of the region's few secular governments and one that Bin Laden had outright condemned. Bush kept his job for that even though the result was to reignite the country's warring factions and incite even more terrorism.
I'm with Thanatos -- the relevant difference (for explaining any intuitive difference) is the fact that the assassination targets specifically identified individuals, whereas the bombing is just indiscriminate killings. This isn't necessarily just a "bias" to be dismissed: there are many cases where the targeted imposition of harms and benefits is clearly worse than a more random distribution (e.g. wrongful discrimination). So we have general grounds to be especially averse to the targeted killings. Still, depending on the precise stakes and consequences, I think it most likely that the indiscriminate bombings would also be impermissible. (It is plainly terrorism, after all.)
this discussion presumes that soldiers are not innocent.
It does no such thing. It presumes that soldiers, whatever their intentions or desires or state of mind, are shooting at you. Whether they are doing so of their own free will, have been forced to do so at gunpoint, or are the victims of orbital mind-control lasers is irrelevant. If they are shooting at you, you can shoot back, even if they are innocent.
This is also why we try, when possible, to let enemy soldiers surrender. If they are willing and able to stop shooting, then we don't want to kill them - regardless of whether they are innocent or seething with hatred for us.
I don't see what changes if we assume that the assassinations option was chosen because the deterrence value of it was greater than the combined deterrence and incapacitation value of the bombings
Deliberate targeting of innocents is widely condemned, even if it is the most effective means to accomplishing an objective. The U.S. could end foreign terrorism much more quickly and cheaply by simply killing all foreigners; the reason we don't do this is not the lack of technical means, but because we find it morally repugnant.
If you have to kill innocents in order to defend yourself, you are in the clear. If you kill innocents because it is the easier way to defend yourself, you are in the wrong.
One can now, of course, begin constructing scenarios where the choice is between spending the entire world's GDP to defend yourself without losing an innocent life versus defending yourself for free at the cost of a single innocent life. I don't believe that such artificial scenarios, hard as they are to solve, refute the fundamental difference between your hypothetical bombing campaign and your hypothetical assassination program.
"You have no control over what family you're born in to, but you at least have some control over how your country turns out."
This doesn't make sense to me. Mightn't it be less costly to disown one's family than to disown one's nationality? And isn't one likely to have very much more control over the state of one's family than the conditions of one's entire country?
A point that I've never seen anyone raise is that this discussion presumes that soldiers are not innocent. It presumes that any young men who are serving in an enemy's army are doing so of their own free will, and because they are bad people.
In many countries during wars, the rates of military service of men in the age group 18-25 approaches 100%. To suppose that this entire demographic is less innocent, and more deserving of death, than, let's say, women ages 30-40, is absurd.
"Ian, bear in mind that the Taliban suppressed dissent. Do you suppose that every single German alive in 1938 was a Nazi?"
Ben, I think the majority of Germans were probably not active believers, just passive followers keeping their heads down and trying to survive.
But if one's survival strategy is to try and not be noticed, to make yourself irrelevant, you have to understand it might not be a perfect strategy. By making yourself irrelevant to the activities of the dictatorship, you also by definition make yourself irrelevant to anyone trying to stop the dictatorship.
So whereas they might give warning and consideration to the resistance, or actively target the true believers, they will not factor you in to their calculations either way. In a world where other countries are likely to come and stomp your misbehaving government, passivity is no longer the safest option, resistance is.
I agree with Psychohistoran. Moreover, the second case could as well be "Let's kill five hundred Koreans" - there is no purpose in killing those who are not the enemy.
This really isn't a complicated problem. Civilian casualties as the result of an attack on a military target is considered acceptable. Civilian casualties caused for the sake of causing civilian casualties are not acceptable.
If the original hypothetical had some military purpose (say, we know 10 of these 510 family members are in fact top ranking al-Qaeda operatives, but we don't know which ones), then there would be a meaningful comparison between the two hypotheticals. As it stands, there does not appear to be much of a contradiction.
One can now, of course, begin constructing scenarios where the choice is between spending the entire world's GDP to defend yourself without losing an innocent life versus defending yourself for free at the cost of a single innocent life. I
That scenario will be difficult to construct. As soon as costs begin to get large they become equivalent to some number of innocent lives. Money represents resources, and as long as there are absolutely poor people, it's possible to spend small amounts of money in ways that prevent innocent deaths. Food and various public health measures cost money, and people die every day from their lack. Spending outrageous amounts of money on a war carries an opportunity cost in these lives, because even if we did none of these alternate things with the money explicitly, if it were simply not taxed, some small percentage of it would be donated to potentially lifesaving causes. The percentage spent and efficiency do not have to be very good to make a lot of lives saved out of what we've spent on the Iraq war, say. The percentages say that merely letting the taxpayers keep that money would have saved a whole bunch of lives. Actively using it with life-saving in mind would save a ridiculous number of lives.
In other words, the taboo against assassination exists entirely for the benefit of potential victims, who tend to be high-ranking and influential and famous and so able to perpetuate such a taboo.
I think we have a winner. Nice reasoning, Jeff.
Another version would be if the government could design a virus that would be carried asymptomatically by most people, but was designed to kill the host if they were genetically very similar to Osama bin Laden. Assuming such a virus could be made and was guaranteed not to kill anyone who wasn't at least a second cousin of Osama, what would be acceptable casualties?
In any case, I would hope that the modern US public disapproves of 'terror bombing', that is bombing designed primarily to cause human misery rather than to destroy the enemy's means of war. As far as I know the US hasn't used an explicit terror bombing strategy since WWII, though US bombing strategies in Vietnam are debatable on this score.
I always thought the taboo against assassination was because the thought process of military leaders (President/King/Emporer/Tsar/whatever) went something like this:
If we kill ten thousand of their soldiers, they will kill ten thousand of ours. That is an acceptable loss for our noble cause.
If we try to assassinate their leaders, they might try to asassinate me! That is not acceptable under any circumstances.
In other words, the taboo against assassination exists entirely for the benefit of potential victims, who tend to be high-ranking and influential and famous and so able to perpetuate such a taboo.
The best explanation is that we fundamentally tend to view directly intended acts differently from their forseen consequences (remember the Trolley Problem); so "conducting a bombing campaign with the knowledge that 5000 innocents will die" seems to fall under a different moral heading than does "assassinating 500 people".
To distinguish this from the hypothesis that it's about the specificity of the 500 family members, note that we would be just as appalled by an operation that rounded up 500 random Afghanis and killed them as a deterrent to Al Qaeda. The only thing that removes the moral outrage in the first case is our notion that intended acts are morally incommensurable with not-directly-intended consequences- a dangerous bias indeed, since those on the receiving end tend not to see the distinction so clearly.
This is the same exact mistake that Sam Harris makes in his book when he compares bombing casualties to torture. Perhaps they are morally comparable, but we are not judging the actions of individuals. We are judging the actions of governments. A government that can order torture and assassinations is a good deal more dangerous than a government that can order bombings. There are certain lines we must never let our governments cross.
My cynicism is such that I'm not sure that Bush would have been impeached. He'd have caused an almighty economic hoo-ha by murdering 500 members of one of the richest families of the country's leading oil producer, but I think he'd be able to present it as all the fault of the bad guys.
Killing 500 people with the surname Bin Laden would have made more sense than invading a country with no links to Al Qaeda - in fact, one of the region's few secular governments and one that Bin Laden had outright condemned. Bush kept his job for that even though the result was to reignite the country's warring factions and incite even more terrorism.
Wonderful. On a blog called Overcoming Bias, almost all of the commenters defended their intuitions rather than trying to reason logically.
I'm with Thanatos -- the relevant difference (for explaining any intuitive difference) is the fact that the assassination targets specifically identified individuals, whereas the bombing is just indiscriminate killings. This isn't necessarily just a "bias" to be dismissed: there are many cases where the targeted imposition of harms and benefits is clearly worse than a more random distribution (e.g. wrongful discrimination). So we have general grounds to be especially averse to the targeted killings. Still, depending on the precise stakes and consequences, I think it most likely that the indiscriminate bombings would also be impermissible. (It is plainly terrorism, after all.)
Phil Goetz:
this discussion presumes that soldiers are not innocent.
It does no such thing. It presumes that soldiers, whatever their intentions or desires or state of mind, are shooting at you. Whether they are doing so of their own free will, have been forced to do so at gunpoint, or are the victims of orbital mind-control lasers is irrelevant. If they are shooting at you, you can shoot back, even if they are innocent.
This is also why we try, when possible, to let enemy soldiers surrender. If they are willing and able to stop shooting, then we don't want to kill them - regardless of whether they are innocent or seething with hatred for us.
I don't see what changes if we assume that the assassinations option was chosen because the deterrence value of it was greater than the combined deterrence and incapacitation value of the bombings
Deliberate targeting of innocents is widely condemned, even if it is the most effective means to accomplishing an objective. The U.S. could end foreign terrorism much more quickly and cheaply by simply killing all foreigners; the reason we don't do this is not the lack of technical means, but because we find it morally repugnant.
If you have to kill innocents in order to defend yourself, you are in the clear. If you kill innocents because it is the easier way to defend yourself, you are in the wrong.
One can now, of course, begin constructing scenarios where the choice is between spending the entire world's GDP to defend yourself without losing an innocent life versus defending yourself for free at the cost of a single innocent life. I don't believe that such artificial scenarios, hard as they are to solve, refute the fundamental difference between your hypothetical bombing campaign and your hypothetical assassination program.
"And no, striking at the enemy in any way no matter the cost in lives or public perception is not always the best policy."
Why? I mean, going to war means killing people, once you get into it you might as well go for the gold.
"You have no control over what family you're born in to, but you at least have some control over how your country turns out."
This doesn't make sense to me. Mightn't it be less costly to disown one's family than to disown one's nationality? And isn't one likely to have very much more control over the state of one's family than the conditions of one's entire country?
A point that I've never seen anyone raise is that this discussion presumes that soldiers are not innocent. It presumes that any young men who are serving in an enemy's army are doing so of their own free will, and because they are bad people.
In many countries during wars, the rates of military service of men in the age group 18-25 approaches 100%. To suppose that this entire demographic is less innocent, and more deserving of death, than, let's say, women ages 30-40, is absurd.
"Ian, bear in mind that the Taliban suppressed dissent. Do you suppose that every single German alive in 1938 was a Nazi?"
Ben, I think the majority of Germans were probably not active believers, just passive followers keeping their heads down and trying to survive.
But if one's survival strategy is to try and not be noticed, to make yourself irrelevant, you have to understand it might not be a perfect strategy. By making yourself irrelevant to the activities of the dictatorship, you also by definition make yourself irrelevant to anyone trying to stop the dictatorship.
So whereas they might give warning and consideration to the resistance, or actively target the true believers, they will not factor you in to their calculations either way. In a world where other countries are likely to come and stomp your misbehaving government, passivity is no longer the safest option, resistance is.