Brad DeLong's Semi-Daily Journal: A Weblog

Brad, curse you for dragging me away from my beloved grading!

Do the standard examples have any force for you? Suppose I'm a strategic bomber who intends to destroy the enemy's munitions factory, and I foresee that the nearby orphans will be killed. You, on the other hand, are a terror bomber who intends the death of the orphans as away of demoralizing the enemy population. You and I are both connected to the dead orphans, I grant, but some people think we're connected in different ways, and that the differences matter, morally speaking.

One possible way of cashing out the difference is to say that, were I in a possible world where I could destroy the factory without killing the orphans, I would, while you're committed to the orpan-killing in itself.

Now, I don't think I buy any of this, but I do sometimes see the difference between "I see that Y is a consequence of X, and I intentionally do X" and "I intentionally bring about Y." My intuition is that the difference doesn't make a moral difference, not that there's no difference to begin with.

[I do think that the argument that people who like bringing about Y are malignant f***s who it is better to not have around has force. And I agree that people who locate munitions factories near orphanages in the hope of so deterring bombing attacks are scum. But it seems to me that the principal use of double effect is so that the strategic bomber can say, "It's not really my fault the orphans are dead. All I wanted to do was bomb the munitions factory." And I don't like that evasion of moral responsibility.]

Another example is a classic one in Catholic teaching. Suppose a pregnant woman is ill, with a sickness that will kill her. The treatment will kill the fetus. This is acceptable if the death of the fetus is a side effect of the treatment, not the point of the treatment (e.g., OK if the women has cancer and the chemotherapy will kill the fetus, but NOT OK if the women has diabetes and removing the load the fetus is putting on her system is the only way to save her).

Not that I agree with this example, but there is a certain logic to it.

[There is?]

I would read this differently. I would say that a moment of inattention has only one chance in 10,000 of killing a pedestrian, while getting them in your sights and gunning it kills a pedestrian with probability one. I would say that while the pedestrian in the first case is as dead as in the second, and that the driver in the first case did a bad thing, that the principal cause was bad luck...

It is, after all, not foreseeable that a moment of inattention on your part will kill a pedestrian...

One analogy that I've thought about before involves traffic accidents - I think most agree that someone who isn't paying attention while driving and accidentally runs over a pedestrian commits a much smaller moral transgression than someone who spies a pedestrian in mid-crosswalk, points his car at him, and hits the gas pedal, even though the end result is identical.

Perhaps the bombing situation is similar; if you drop enough bombs some are going to go astray and kill innocents, but any particular incidence is not as certain of a thing.

Interestingly, I myself was just tested on the doctrine of double effects this last Tuesday (up here in UC Davis). The final having been over for days, I cannot remember who has done the modern work on the subject (I am between a Religion final and a Greek History final, if that offers any excuse), yet I think it hold some application…

One of the modern understandings of the doctrine of double effect is in order for us to claim to be operating under it, we accept for ourselves a higher level of danger. For instance: suppose there is a munitions plant next to a hospital. Say for every X number of feet lower we fly, we have a 10% greater chance of harming only the munitions plant, but a 5% greater chance of having our own pilot be shot down. To claim to be operating under double effect today, we have to be willing to allow our pilot that greater chance of being shot down. We must, to some degree, care more about the lives of civilians—even enemy civilians—than we do for our own combatants…

This, as rightly noted by a previous commenter, helps distinguish the just from the unjust in war. Since war seems to be an essential part of the Human condition, forcing even it into some sort of moral code might be all we can do to minimize the agony it produces…

Phillippa Foot is one of the most famous modern discussers of double effect, in her article "The Problem of Abortion and the Doctrine of Double Effect." That article introduced the closely related trolley problem--see this good wiki (http://en.wikipedia.org/wiki/Trolley_problem) and this APA-approved parody (http://home.sprynet.com/~owl1/trolley.htm).

Here's a political example--Grover Norquist has apparently said it's OK to keep the AMT in place because it mostly affects people in blue states (http://www.thepoorman.net/archives/003567.html). Now, we could imagine someone--call him Mr. TMA--who decides to leave the AMT in place, looks at the state-by-state breakdown, and realizes that it mostly affects the blue states. Mr. TMA foresees the consequence that leaving the AMT in place will disproportionately hurt blue staters, but that's not his purpose in leaving the AMT in place.

But the fact that Norquist seems to think that the AMT should be left in place because it punishes blue-staters indicates that he intends that punishment. And I think that makes us think the worse of him (should that be possible). We judge Norquist much more harshly than we'd judge TMA.

Sorry for political trolling on a philosophy thread, but I thought an example that stirred your juices up might be the best way to bring the intended distinction home!

One thing that strikes me--I'm sure this has been discussed in the literature if it's at all promising--is that the difference between an intended consequence and a foreseen one may come out in counterfactual situations. If the AMT fell mostly on red states, TMA would still want to hold it in place and Norquist wouldn't. That shows that Norquist intends the AMT to punish blue staters and TMA merely foresees that it will hurt blue staters, without intending it to.

Curse all of you, I've just spent the last two hours mulling this over and now I can't get it out of my head.

Ultimately I think I'm in the skeptic camp. I think the doctrine has some intuitive logic but it doesn't do the work it wants to in the hard cases.

I have no problem with the example that FL gave of the bomber pilot vs the terrorist bomber but I am confused by the following:

"It is claimed that sometimes it is permissible to cause such a harm as a side effect (or “double effect”) of bringing about a good result even though it would not be permissible to cause such a harm as a means to bringing about the same good end."

I just don't get this. I think if it's moral to perform a life-saving operation on a pregnant woman that will also kill a fetus I can't understand why it wouldn't be moral if the fetus is, itself, the life-threatening entity.

I don't see how the inclusion of a third party (the illness that is separate from the pregnancy) changes the calculation.

Take this example:

"It would be wrong to throw someone into the path of a runaway trolley in order to stop it and keep it from hitting five people on the track ahead; that would involve intending harm to the one as a means of saving the five. But it would be permissible to divert a runaway trolley onto a track holding one and away from a track holding five: in that case one foresees the death of the one as a side effect of saving the five but one does not intend it."

I really can't see the moral difference between the two situations. In both cases there is an initial cicumstance in which there are 5 people in danger and one person who is in no danger at all. By your action you can change the situation so that the 5 people are saved and the one innocent bystander becomes the victim.

Broadly caricatured I think this is the old tension between utilitatianism's moral calculus and the Kantian "kingdom of ends." We look for ways to resolve that tension because, intuitively we're all utilitiarians sometimes and Kantians sometimes. But I think if we're going to stradle that line we can't do so without accepting that we're occasionally utilitarians.

Go back to the trolly example, the only difference between pushing the person onto the tracks, and using a switch to re-direct the trolly is that the second involves using a tool to act indirectly. It certainly changes our moral intuition, but I don't think it changes the underlying moral situation.

Ha! I guess it is because of this and other such nonsense that the Orthodox church does not pay much attention to these guys.

Like Augustine's belief that angels materialize and dematerialize at will. Yeah, if you believe that there are some people in white robes that would like to have a chat with you.

(I'm gonna duck and hide now - bye!)

This "grading" thing is an interesting dialect difference between Americans and Canadian. In Canada, it is always called "marking." You give marks. You take away marks. At the end, you get a mark on the course, after the professor as marked the exam and the assignments. That marks might be a percent mark, or it may be a "letter grade". That appears to be the main use of "grade" in Canada.

Double Effect?

http://fafblog.blogspot.com/

December 16, 2004

Social Security Solution!

As everyone who has followed Giblets's 50-part series "Social
Security: ARMAGEDDON!"1 knows, Social Security is going to EXPLODE! -
MAYBE! - in FIFTY YEARS! - because it will run out of money, and the
only thing to do is to borrow two trillion dollars from the Mystical
Realm of Faerie to save us from going into debt.

But there may be another solution! Right now the usual bunch of
whiners (old people, sick people, poor people - really, what do they
do except whine?) are whining about Bush's plan to get rid of tax
breaks for health insurance. This could be the Social Security
solution we've all been waiting for! If thousands of companies get
punished for giving their workers health insurance, well damn!
Millions could end up LOSING their health insurance. And the less
health coverage you have, the sicker you are, and the sicker you are,
the faster you die, and the faster you die, the less Giblets has to
pay for your stupid Social Security!

Between gutting health coverage and sending old people off to war,
we've got a great start going putting a dent in American life
expectancy. Right now it's somewhere around 77.2 years. That means
12.2 years where our parents and grandparents can leech off our
hard-earned cash! If we work hard we can push that way lower - down to
75, 70 years, down to 68 and lower if we really work at it, and then
we could just raise the retirement age and not have to worry about
Social Security at all!

So there you have it. Giblets has solved all your fiscal problems at
once! Social Security is saved! The budget is saved!2 More cannon
fodder for Iraq! Three birds, one stone, everybody happy! Except for
the dead people, and hey, Giblets doesn't hear them complainin'! You
can thank Giblets anytime.

1. soon to be a major motion picture by Jerry Bruckheimer
2. at least until the tax cuts become permanent

Cory et al. match my intuitions better than FL (there's a difference but not a moral difference) or DeLong. If you could take steps to prevent or reduce the probability of or the acuity of or to somehow remediate the bad side-effect and didn't, then you'd be culpable -- maybe as culpable as if the plan inherently required the bad effect -- but it would be a different story (to me at least) if you did take those steps.

Well, I'm studying for my law school finals so I'll throw in my two cents to procrastinate.

The Doctrine of Double Effect seems to be describing two of the four levels of mens rea in criminal law, and we're arguing about which two. The four degrees of culpability are negligence, recklessness, knowledge, and purpose. (For good descriptions, see: http://en.wikipedia.org/wiki/Mens_rea.) In his comment, Brad describes the docrtine as negligence (maybe recklessness?) and purpose, but Fontana sees it as knowledge and purpose in his comment.

So, I guess the point is, the Doctrine of Double Effect isn't complicated enough. I think the law does a better a job of parsing out degrees of culpability. But it is based on the same fundamental principal, that even though the pedestrian or the orphans are harmed equally both cases, the degree of moral culpability (and the degree to which the perpetrator had a guilty mind) is different.

This may all just reveal how much of Criminal Law I've forgotten. Oh well, back to the First Amendment...

Is the Double Effect in any way connected to trading away Hudson and Mulder on the same weekend?

There is also the limited knowledge effect for this doctrine. You don't know if it's an ammunition plant at all, or how likely you are to hit it, or how likely you are to hit anything at all.
Consider the real world application of Bomber Command of World War II. The British were so bad at navigation that not only did they fail to disrupt German industrial production, but they killed so few civilians that it was a net casualty loss to Britain. They lost more aircrew than they killed Germans.
Not killing innocent people is just a good rule of thumb.

"The British were so bad at navigation that ... they killed so few civilians that it was a net casualty loss to Britain. They lost more aircrew than they killed Germans."

Having grown up in Hamburg when it was still scarred by the ruins of WWII, I rather doubt this statement. What is your source for it ? I just might believe it for the earliest bomber raids, but not for the whole war.

This looks like a stupid idea to me, and if respected moralists take it seriously then maybe that indicates that thinking too much about morality tends to make you stupid.

It looks like the general case is that you do an action that has two parallel predictable results, A and B. A is generally considered good while B is considered bad, and there's every reason to expect both. So if you do the action and you *want* result A while you *don't want* result B, then you're good but if you want B then you aren't good.

I just don't get it. If a bad guy is threatening an innocent victim and I repeatedly stick a knife into him until he dies, then I'm good if I want to save the victim but I'm bad if I enjoy killing the victimiser?

What if a surgeon enjoys cutting people; he fantasises about killing them but actually does a skillful job of surgery. Is he guilty like the one who enjoys actually killing people?

What is the point of this whole exercise anyway? It clearly doesn't make sense to pay a lot of attention to the intentions of felons while arriving at sentences. That just encourages them to lie about their intentions.

Or does that make sense? Maybe if they lie about it, they'll start to become what they claim they already are? Maybe if they have no reason not to lie they might brag about their exploits and influence others to copy them? Maybe the soviets were right that there's social value in getting criminals to confess and recant....

But apart from the pragmatic value in bribing bad guys to claim their intentions were good, what use is a morality that depends mostly on what happens inside people's heads?

Try this one. You're taking an ethics class and the grades will be on a curve. Faced with a question about double effect, you cheat. You have strong reason to believe that this cheating will raise your standing from 2nd place to 1st. The result of your cheating will be good -- your grade will go up -- and bad, everybody else's grades will go down. If you want your grade to go up but you think it's a regrettable side-effect that the rest of the class is harmed, then double effect works in your favor. And in that case if you do after all get caught the instructor should let you off because your intentions were good.

Have I seen the absurdity of it, or is there some actual point to this exercise that I've completely failed to grasp?

Wait, I see something. It was OK to kill a fetus as a side effect of saving the mother, but not when killing the fetus was *required* to save the mother. Let's try that one with the orphans.

If the orphans just happen to be living beside the munitions factory then it's double effect; the bomber wants to bomb the factory and doesn't want to bomb the orphans though he knows he'll do both. But if the orphans are the ones who work in the munitions plant, and if they survive they'll rebuild the plant and keep making munitions and so killing them is central to the plan, then double effect does *not* enter into it and it's just wrong to kill them.

OK, I think I've got it. This is a sort of christian zen. They give you koans like this and it's supposed to get you to give up on logic and accept the ineffability of god. Probably they closely observe the trainees and when they see one who continues to argue about it with the intention of putting logic above faith they figure he's on the wrong side of the double-effect boundary and they hit him with a stick. To wake him up.

That must be it. Christian zen.

xian zen...worked for phil jackson.

“We must, to some degree, care more about the lives of civilians—even enemy civilians—than we do for our own combatants…”

I’m not sure I would want you as my commanding officer. The first duty of a government is to protect it’s own citizens, including the military, over the interests of the enemy, including enemy civilians. Especially when the enemy civilians are an integral part of the enemy’s war effort. Of course the wanton destruction of the enemy whether military or civilian is a war crime. For example, the bombing of Dresden in WWII would seem to qualify as a war crime, but as the historian Martin Gilbert pointed out, you have to blame Stalin. Churchill was furious when informed of Dresden firestorm, that action ran contrary to his expressed policy. Stalin had requested the bombing, falsely claiming it was a military necessity, and Churchill didn’t find out until after the fact.

Does the double effect relate to the principle of self-defense? If a robber tries to steal from me on the street, when can I use lethal force to stop him? According to the law in most states, only when a “reasonable” person would judge that his own life was in imminent danger. You may not use lethal force to simply protect your property. However suppose your property is a frozen organ that you are delivering to a hospital to save someone’s life, and you are afraid the robber will destroy it. The law might not take your side, especially if the robber was from a protected minority group and the patient wasn’t. Could you invoke the double effect as a defense?

(How do I get paragraph spacing?)

I have come to the conclusion, after long discussions with Catholic ethicists,that the chief purpose of th edoctrine of the double effect is as a sort of intelligence test. It allows the Church to have a doctrine which is in practice flexible, but which sounds reassuringly solid to people too stupid to grasp even the simplest form of the doctrine. In other words, it lets you should very loudly that you are upholding the principle around which you are working. Examples here include -- obviously contraception (in Britain at least this is no problem if you do it for the sake of your marriage) but also euthanasia, where the "relief of suffering" allows catholic doctors to give injections that they know -- if they are competent to practice at all -- will kill the patient, since they are simply trying to relieve her pain.

In other words, it has great social use,especially for Straussians, and very limited philosophical worth.

Ooh, I think I see a deep connection here between the "acceptability" of some courses of action according to the law of double effect and the usual findings in so-called Framing Effect experiments by Kahneman and Tversky. I'd spell these out for you, but I have to, uh, do a whole bunch of grading today. Here's a pair of stories in the mean time:

You are a doctor treating an epidemic in a small village of 600 people; all will die unless you act. You have two vaccines you could use to try to help the situation. Vaccine A will save 200 lives in this situation. For Vaccine B, there is a 33% chance that everybody will be saved, but a 66% chance that all 600 will die. What do you choose?

OK, got that? Let's say you chose correctly and now you get promoted:

Now you are a doctor treating a different epidemic in a small village of 600 people; all will die unless you act. You have two vaccines you could use to try to help the situation. If vaccine C is used, 400 people will die. However, if vaccine D is used, there is a 33% chance that nobody will die, and a 66% chance that all 600 people will die.
What do you choose?

From what I can tell reading this thread, the Church does not council indifference in both cases (which are the same case except for the framing) while Brad deLong and other card-carrying economists, who gave their highest prize to Kahneman (in part) for this finding, would point out that all 4 choices are identical.

So, do I get any points for this? :-)

Okay, since I am neither an economist nor a theologian, I will try to translate one of these unworldly belief systems into the other.

Double intent, in economic terms, seems to go to preferences. Let's take a gross case, and pretend I want to fire a gun at a target. Let us also pretend that I am completely indifferent to whether a person is standing in front of the target. In this case, the person has no marginal effect on my preferences for shooting the gun at a particular moment.

Is this act morally any different than deliberate murder? I don't know, but it is economically distinguishable. The law also draws a distinction. It is not "intentional" murder, but rather "depraved indifference" murder, which in some states might draw a lesser penalty.

I share Brad's puzzlement on the moral distinction drawn by double intent, but find the idea easy to understand and occasionally even practical.

Consider the law of market manipulation. It requires "specific intent"--which means that a "double effect" defense works. Let's pretend that you wanted to buy a huge long position in the slunk futures market, because you thought that slunks would appreciate in value. (You don't want to know what a slunk is, and don't need to know.) Your long slunk position would not be market manipulation, even though you were also pretty sure that, in the course of building your position, you would corner the market. Only if you bought your slunks or long slunk futures with the intention of cornering the market, you would be a market manipulator.

The double intent defense actually make sense here, although people might draw the lines differently than the law does. We want people to invest aggressively in futures markets--the better for price discovery. We also don't like corners. The current rules protect everybody except those whose actions are inexplicable as aiding price discovery. For example, pretend our slunk merchant secretly leased all the slunk rail cars, parked them somewhere, and then went long. There is no non-manipulative way to explain this course of conduct. The double intent defense does not wash. However, it is easy to construct a double intent defense if the long's activities are limited to buying and selling. Of course, lying is easy and it is normally hard to falsify a lie about mental states. But guess what--it is very difficult to corner a market by merely buying and selling. Things kind of work out.

He hoped and prayed that there wasn't an afterlife. Then he realized there was a contradiction involved here and merely hoped that there wasn't an afterlife.

"To claim to be operating under double effect today, we have to be willing to allow our pilot that greater chance of being shot down. We must, to some degree, care more about the lives of civilians—even enemy civilians—than we do for our own combatants…"
That's how it should be in Iraq. We chose to invade. The welfare of the country is in our hands. It seems to me the welfare of our troops is of a secondary moral concern to our duty to provide for the welfare of the Iraqi people.

It seems pretty sensible to me, the distinction between killing someone in self defense and committing a murder. But, I can certainly see why it's fuzzily worded enough to be of "limited philosophical worth" as Andrew Brown says. It's one of those truisms that can be better examined in other ways.

Maybe it's good to look at alternatives.

If you don't bomb the munitions plant, what's the likely outcome of that? If it's unacceptable to leave the plant unbombed, then there still might be a third choice. Maybe you could airdrop a suicide mission ahead of time, whose job is to get the orphans away before the bombing and then surrender -- or go back into the ruins and sabotage them some more.

If the alternative to killing somebody is being killed yourself, that's a choice that's worth considering carefully. But what about the difference between killing somebody for the $20 they're carrying versus killing them to keep them from taking $20 from you? Either way it's killing somebody for $20. There's a difference -- it's your $20 versus their $20. But....

After thinking about it for a long time I have come to believe that the doctrine of double effect operates as a way of maintaining that one has upheld one's principles (e.g., we don't kill orphans or fetuses) while nonetheless achieving one's "greater" aims, however defined. It's a sophisticated way of absolving one's self of moral responsibility for one's actions, and perhaps from a theological perspective, from the accusation of being a utilitarian.

Consider the true Catholic doctrine on ectopic pregnancy:

Okay -- Remove the mother's tube. The operation doesn't "actively" kill the fetus, though that's the inevitable result.

Not okay -- Give the mother methatrexate, which stops the pregnancy and usually causes a miscarriage to follow. This action "intends" the death of the fetus.

In either case, the mother would do everything possible to save the fetus if she could, but she can't. Of course, the "moral" method destroys her fertility, but what's that compared to an abstract principle.

Jonathan, you're assuming the independence axiom here. (Channeling Mark Machina). People's preferences over the lottery depend on who else survives. Maybe without others around, people would weight the probability of their own death less severely--nobody to enjoy life with. Or, alternatively, there might be some sort of benefit to being a survivor, like after the Great Plague in Europe. Populate the world with my progeny, that sort of thing. That will affect the preferences over the two lotteries. There are other issues with the timing of uncertainty vs. risk that I won't go into here, but there's no reason to expect a clean representation of a "social planner's" utility over lives saved.

This, by the way, is the major failing of expected utility theory in my opinion. To get a proper representation of things, we have to specify weights on EVERY state, and monotonicity is no longer a useful property. So that handles the difference between A and B, and C and D. There are real differences there.

You do correctly point out that people have preferences over A vs C and B vs D. This is the real framing effect and it's pretty big. Saving 200 lives and letting 400 people die are identical outcomes. But people don't naturally think clearly in terms of consequences. They think in terms of motives, as this entire thread suggests. No need to rehash the Kling vs Krugman arguments, but this is a real problem that needs serious discussion.

How many people distrust free markets because of the profit motive, despite their generally beneficial consequences? How many people approve or disapprove of the Iraq war because of its motivations, regardless of the positive or negative consequences? How many people's opinions of the minimum wage or welfare reform do not involve the incentives that poor families face? Ideas and values matter. This is difficult to model.

Off topic but it might actually help explain:

In criminal law, there are four categories of mens rea, or culpable intent, behind a criminal act. Let's just use homicide since it's easiest:

1) you intentionally kill someone--your actions will result in their death and you're happy about it.
2) you knowingly kill someone--you know your actions will result in the death of another, but you would just as soon they not die; you will not be disappointed if they survive.
3) you recklessly kill someone--you know there is a substantial risk that your actions will result in another person's death.
4) you negligently kill someone--you should have known, but did not realize, that there is a substantial risk your actions would result in someone's death.

Say you're an assassin/terrorist who wants to kill the dictator of Blogtopia. You wire a car bomb to his car, to ignite when the car is turned on. You are certain that his driver will also be blown up by the bomb, because the dicator is physically unable to drive himself. You feel bad about this, you have nothing against the drive, but you go ahead with the plan anyway. You also realize that the ambassador's wife may be in the car with him when the bomb goes off. You think those are the only possible innocents who will die as a result of your actions. As it turns out, you're wrong; shrapnel from the blast kills four more innocent bystanders.

You have intentionally killed the dictator, knowingly killed his driver, recklessly killed his wife, and negligently killed the bystanders.

As I understand it, double effect doctrine is drawing the line between a KNOWING killing and an INTENTIONAL killing. It's counterintuitive, because our laws almost never draw the line there--it's always between innocent and negligent, negligent and reckless, or reckless and knowing, or all of the above.

Or, to be less boring about it, from the Brothers Karamazov, the chapter about the Grand Inquisitor:

"Rebellion? I am sorry you call it that," said Ivan earnestly. "One can hardly live in rebellion, and I want to live. Tell me yourself, I challenge your answer. Imagine that you are creating a fabric of human destiny with the object of making men happy in the end, giving them peace and rest at last, but that it was essential and inevitable to torture to death only one tiny creature- that baby beating its breast with its fist, for instance- and to found that edifice on its unavenged tears, would you consent to be the architect on those conditions? Tell me, and tell the truth."

"No, I wouldn't consent," said Alyosha softly."

The torture of the child in Ivan's hypothetical fails the doctrine of double effect: you are instrumentally using the torture of a child to achieve a good end.

But say, instead of this, you were to give mankind free will, knowing with 100% certainty that some men would exercise this free will to torture innocent children. That passes the doctrine of double effect.

caveat: I'm not a philosopher and haven't taken any philosophy classes, so I could easily have this wrong. Do I pass the test FL?