Yesterday’s nightmare scenarios triggered some good discussion, so let me throw out another one, which I think will help to isolate some of the issues that came up yesterday. Sometime next week I’ll try to summarize the best of the comments and ponder what we’ve learned.
|
In front of you are two childless married couples. For some reason, it’s imperative that you kill two of the four people. Your choices are: A. Kill one randomly chosen member from each couple. All four people agree that if they die, they want to be well remembered. Therefore all four ask you, please, to choose A so that anyone who dies will be remembered by a loving spouse. If you care about the four people in front of you, what should you do? |
Argument 1. For goodness’s sake, they’ve told you what to do. If you care about them, of course you should respect their wishes. Choose A.
Argument 2.Once the killings are over, Option A leaves two grieving spouses, whereas Option B leaves one relieved couple. Surely two dead plus two happy is better than two dead plus two sad. Choose B.
Which argument do you buy? And what’s wrong with the other one?
Do they all agree individually or does each couple agree jointly? If all individually then they might impose a cost on their spouse without taking that cost into account.
In the case of the former I’d go for B. Cost-benefit does not include the benefit to the death spouse, the benefit is non-existent and so is he/she, whilst the cost to the surviving spouse remains.
This might change though if I’d have to announce the decision beforehand, then the benefit of anticipating being well-remembered is there versus the expected cost of the grief.
In the case of the latter I’d go for A, I am not willing to second-guess the couples joint decision unless given reason to.
it should read ‘benefit to the dead spouse’.
Argument B ignores the wishes of the couple that ends up dying. Clearly, in practice, people do care about what happens after they die – otherwise nobody would write a will, nobody would have even suggested the idea. There’d be no debate about global warming, no epic sci-fi, no religion. Since the decision is made before they die, it is incorrect to ignore their utility function at the point of the decision. Argument B is fallacious.
There are only 4 people affected by the decision. All 4 of them agree about what the decision should be. Therefore Argument 2 is irrelevant.
If you care about the four people in front of you, you should ignore their wishes because you understand that the horror of the moment robs them of the careful reflection that their situation requires. The cost of Choice A includes the possibility of losing two prospective mothers and the offspring that they might create. Choice B preserves a couple that has the potential to procreate and a higher likelihood of being productive members of society than the bereft survivors under Choice A. Argument 2, therefore, prevails
I’ve seen argument 2 before. That’s the antiabortion “Women can’t
possibly have enough brains to make an intelligent decision” one.
Therefore, we must impose enough obstacles to make them think.
Waiting periods and medically unnecessary sonograms might help.
In this particular case, four adults have each expressed their
definite and unanimous choices. It’s the height of arrogance to use
argument 2 to abrogate their choice because, of course, you know
better.
What if the couples are swingers?
Mike H says that if people didn’t care what happened after they died, “There’d be no debate about global warming, no epic sci-fi, no religion.” I think he misses the point of both epic sci-fi and religion, and possibly of the global warming debate as well. Authors don’t write sci-fi for future generations, they write it for the present generation, even if there is a benefit to future generations. Believers in religion generally don’t follow religion because they want to escape hell; they follow because they believe their religion is the best way to live even here on earth. And if you ask global warming activists, they will usually tell you they’re worried about climate change within their own lifetime, justified or not.
I would kill one at random, then allow the remaining three to revise their priors before killing the second. To avoid moral hazard problems, I would insist on unanimity (or ask for a volunteer).
Argument 2 just doesn’t work for me. The couples have revealed their preferences to you and have stated that 2 dead and 2 sad IS better than 2 dead and 2 happy.
Kill a couple and build a large memorial honoring them. That will cause them to be remembered far more than the remaining spouse.
Now that we’re talking about offing people for the greater good I kinda wish Yoram Baumann were here to call it grinchy.
@Scott, Mike H: Mike H is clearly correct that people care about what happens after they die. As I put it on the previous thread, old men plant trees they will never see grow.
I’m not sure fallacious is the word for argument B. Arrogant perhaps. It is the statist’s answer: I know better than you do. But it’s not fallacious because expressed and revealed preferences can differ, sometimes others do know better. I have little trouble making decisions over the objects of 3 year olds.
Kill a couple.
Yep, living people care what happens after they die, and yes it’s usually good to respect peoples’ expressed wishes. But by killing them we’ve removed their concerns for the future and their wishes.
It’s not like abortion. Prevent a woman from having one, and she may very well wish for years and years and years that she hadn’t been so prevented. Ignore these peoples’ desires and you’ve guaranteed that there won’t be any similar wish.
Take a couple and put them in separate rooms. Tell each that his/her spouse will survive to remember. Then kill both of them and let the other couple live.
Isidor and Ida Straus, where are you when we need you?
I buy argument one because I believe that other people know their own preferences better than I do.
If I understand Landsburg’s point on this one, I think what he’s driving at is that some of us think we know better about what’s best for people than the people in question do themselves. But I don’t profess to know what another person’s preference is, and I am wont to take people at their word.
So Option A is the only one that makes any sense to me.
This, of course, is assuming I play along with a game that requires me to kill people.
On hypotheticals in general:
I often have difficulty appreciating whether the questioner wants me to actually project myself into a described circumstance and provide a visceral reaction, or wants me to wrestle with an abstract intellectual exercise. Would I pull a lever to cause a trolley to shift tracks and kill one person rather than 10? Would I push a person over a railing to achieve the same result? The answer on a visceral level is no and no; almost certainly I’d be paralyzed with alarm under either circumstance and would be too distrustful of my understanding of the circumstances to take action. But I suspect that this answer misses the point of the hypothetical.
On the current hypothetical:
1. To what extent do we honor the preferences of people who are dead relative to the preferences of people who are living? To what extent do we honor the preferences of people who are soon to be dead relative to the preferences of those who will survive them?
2. To what extent do people’s statements of their preferences reflect what they will actually value? The field of affective forecasting suggests that people misapprehend the things that will make them happy – and do so in predictable ways.
3. Assuming I conclude that people are mistaken about what is in their own long-term self-interest, should it matter? If I want to promote someone’s welfare, should I rely on my own judgment about what is in their best interest or on theirs? Is a bias toward relying on my own judgment merely hubris? Is a bias toward relying on other people’s judgments merely an effort to shirk responsibility?
The conundrum is phrased as a one-off. I think that unfortunately means it cannot be used as a test for how people would think about policy decisions, which are not one-offs.
The key difference is that in this case, we could ignore the wishes of the two people who die, and they’d never know. We can debate whether that’s unethical, but they wouldn’t be around to be unhappy about it.
By contrast, if you have a society-wide policy of ignoring the wishes of people after they die, that will make people unhappy whilel they’re still alive.
@Bennett Haselton: very good point. Plus it also means one’s answer to the one-off could be dicey too. Lots of research shows that people implicitly apply repeated-game logic to one-time games. It’s almost impossible not to.
Off with their heads!
@ Bennett and Ken B — even assuming this isn’t a one-off…
We know dead people have been voting for Democrats since at least the Eisenhower administration. However, I haven’t found anyone who really defends the practice. Why expand the right of the dead to vote now? How many things might be different (worse?) today if we made societal decisions based on what the dead wished?
Life is for the living. Option B is definitely on the table.
@Scott H:
B is definetely on the table. I just said it’s arrogant, not based on a fallacy. That means I consider it a legit option.
There’s a difference bewteen voting and having your previous vote count.
” How many things might be different (worse?) today if we made societal decisions based on what the dead wished?”
We might follow the constitution.
Ha!
I visited my brother in Chicago during an election season. An TV ad aired that panned across a military cemetery as a voiceover intoned, “These people died defending your rights – including your right to vote. So this Election Day….” My brother turned off the volume and provided the rest of the voiceover: “This Election Day, all these people will be voting; shouldn’t you?”
We’d be able to make private recordings all the live concerts we attended?
Kill them all and let God sort it out?
@mobile: Oh, we want you over at Murphy’s blog!
I need to make the assumption that I think is reasonable but is unstated, which is that all four people are of appropriate age, sound mental condition, and fully informed about their situation and the options. Under that assumption I’m likely to respect their wishes.
I believe that this is another case where the specific will not generalize well, but individually I can be assured of fully informed consent and can act on that basis, I think. Unsurprisingly I also support right-to-die legislation that comes with what I think are adequate informed-consent safeguards.
It’s often tricky to decide when responding to the exact literal words of a problem like this is missing the point. In this case, it looks to me that taking future children into consideration is like responding to “can God make a rock so big he can’t lift it” with comments about gravity–sure, the question is worded so as to allow it, but that’s not really what it’s asking, and you’re just nitpicking the wording.
Comparing only the things we were meant to compare, I’ll have to give an answer that has been echoed in various forms by other posters: argument 1 states that a dead person who is remembered is preferable to a dead person who is not remembered; but argument 2 states that all dead people are equally preferable. The contradiction boils down to that.
Which is worse, killing a person swiftly and painlessly, or killing a person by long and painful torture? By the same reasoning used in argument 2, both actions will result in one dead person, who doesn’t care (being dead) how he got that way, so both are equally preferable. They shouldn’t be equally preferable, but there are several ways to *phrase* the difference. For instance, you can count the torture as negative utility for the dead person and reject the assumption that there can be no utility for a dead person. Or you can count the torture as negative utility for the living person, and reject the assumption that utility has to be measured after the fact. Or you can say that “being in a state which was reached via torture” itself has negative utility.
It’s basically an accounting trick to decide exactly *how* you’re going to count the torture. But you have to count it *somehow*. The same goes for the original problem: there are several different ways you can consider “it’s better for a person to be remembered”; but you can’t just ignore it as argument 2 does.
Incidentally, my answer to the trolley problem is that in the problem as given, I would have to kill one person to save 10, but I wouldn’t do that in the real world since the assumptions of the problem–that I can know absolutely that those are the only choices and what each outcome is–would never be true in reality.
@nobody
“Assuming I conclude that people are mistaken about what is in their own long-term self-interest, should it matter?”
My answer is similar to the trolley problem: If I can be absolutely sure that someone is mistaken, you can make a case for overriding their wishes. But I can never be absolutely sure, and given the nature of human errors, some of the worst things in history have been done by people who think they know enough to impose decisions on others against their will–in other words, this sort of error is especially pernicious and we should take special steps to avoid it.
Even in the rare circumstances where I can be almost absolutely sure, it’s impossible, as a practical matter, to make a policy that would take effect only in such rare circumstances.
Bennett Haselton has convinced me.
Bennett cares about what happens after his death, and when his life is in our hands. In order to avoid distorting his choices we want to convince him his wishes will be seriously considered. The only plausible way we can do that is to count heavily everyone’s wishes in such situations. We will prefer a society that does that. That society picks A. Especially as we have consulted with all those most directly affected.
In yesterdays’s dilemma 1, most people place a serious value on the future of the species (not Peter Singer). Again we want to respect their wishes as much as possible. That society chooses 7B flips.
Memory, shmemory. The answer is B. The pain of loosing a spouse won’t be consoled by the thought that he/she is remembered.
@ Ken B
Landsburg is making me kill these people so my opinion — or any other killer’s opinion — needs to count at least as much as the folks getting killed. The other two people and I will have to continue living in this evil Landsburg dictatorship we’ve hypothesized.
After the fact, the only way the living couple could argue with an option “b” decision would be to say that they wish their spouse was now dead. To which I would reply “Boy, I wish you would have mentioned that a little earlier!”.
Scott H hit the nail on the head!
Most people here are missing the point entirely. Of course the only reasonable answer (if you care about them, if you don’t maybe A will make you feel less guilty) is B.
Dead people don’t have utility functions, so you only consider the utility of the living ones.
Of course, I assume that is the answer Steven expects, because one can argue that if dead people don’t have utility functions, neither do unborn people, but I personally consider that not obviously true.
It’s not the case that this is about people knowing their own preference than someone else knowing their ‘true’ preference.
The people who wish to be remembered are imposing their preference on their partner. It clearly says they would like to be remembered, not, that they would like to be the one remembered. It’s shocking no one has pointed this out.
Therefore, it is not the case that we know the utility before the action. The grief of the living spouse could, and I suspect, would, negate any benefit to he who gets remembered.
It’s like the problem of average number of students a student thinks is in an average class vs average number a professor thinks is in an average class. More people see the classes with more people. We typically call this a bias in observer but, here, that bias is worth something in that both living people experience additive benefit, whereas a living/dead combo experience subtractive benefits.
The answer, undoubtedly, is B.
@Scott H, Ken…. this reminds me of a vignette I read that put various moral dilemmas like this in front of the protagonist. The last one was something like “I’ve already made my choice, and done the deed. Which of A or B do you hope I did?”
@Scott : there would be no epic sci-fi, because there would be no market for speculation about the distant future. People wouldn’t buy Foundation or Dune or similar if they didn’t care about what happened after they died. I’ll partially concede your point about global warming, say, 30% of it.
I’m sure the UofR president will be relieved to see Steve has stopped talking about sex and is now blogging about the optimal way to kill married people.
Whatever the unstated rationale, as the killer I should accept moral responsibility for the choice. Therefore I should select option B which preserves an intact family. Option A leaves both survivors with the guilt of having been complicit in the death of their spouse.
The only justifications for my choosing option A would be 1) confidence in an afterlife (reducing the absolute character of death) or 2) altruism in the survivors so strong that they will be miserable knowing that someone else has been “forgotten” on their account.
You should turn it into a reality TV show and have the public vote on who should live and who should die.
You could have teams of moral philosophers arguing for the two options as well as the 4 people putting forward their views.
Week 1 ends with a vote on the first person killed
Week 2 ends up with a vote on the second person killed
“After the fact, the only way the living couple could argue with an option “b” decision would be to say that they wish their spouse was now dead.”
By the same reasoning, if the choice is between torturing person A to death or killing B painlessly, if I decide that A is to be tortured, B can only argue against it by saying he’d rather be dead himself.
Your response implicitly assumes that benefits only accrue to people who are alive and capable of arguing for other options. By your reasoning, harm to anyone who is then killed off (and therefore cannot argue for another option) doesn’t count. This is one of the issues in question, so you can;t just assume it to be true.
Is this hypothetical or are you asking what I wish I had done the other day… ooops… I didn’t say that!
WWMD: What would Machiavelli do?
“The new ruler must determine all the injuries that he will need to inflict. He must inflict them once and for all.”
“If an injury has to be done to a man it should be so severe that his vengeance need not be feared.”
“Hatred is gained as much by good works as by evil.”
In short, regardless of what people say, if I’m the instrument by which their beloved spouse dies, I’m going to earn their hatred. Option A enables me to earn two healthy, deeply aggrieved enemies. Option B enables me to earn two healthy, deeply grateful friends. Just sayin’….
nobody-really gets my vote. The preferences of the guy making the decision are what’s relevant.
Suppose I have a choice between sending my son to Harvard or some for-profit Scam College. Scam College really wants my money, Harvard really doesn’t want it because my son is an idiot. Scam College President gets to pay alimony and child-support and pay for hair plugs if he gets my money. Harvard gains nothing and, in any case, my son is an idiot who will never move out of my basement or get a job which doesn’t involve wearing a cap.
I’d still send my son to Harvard coz that’s my preference.
The only situation where that would not be true is if I were an Agent not a Principal or had a legal duty of care or there was some other factor militating for heteronomy.
Let us suppose I have no preferences one way or the other. Then I should toss a coin. Indeed, tossing a coin is a way of signalling that whatever is decided, you acted without malice and this should mitigate vengeance by an aggrieved party.
The deontology/consequentialism debate is played out. We now know that not only is information costly but processing it is costly and subject to concurrency problems. There is no deontic which does not have hysteresis such that its results are consistent. So
toss a coin or act according to your own preferences.
Sam Wilson seems to have hit on something. Try this scenario:
You are presented with two people, one of whom has just had their spouse murdered by an evil madman: the other two are a happily married, childless couple. You have the choice to either kill the widow(er) or kill one of the happily married couple at random, which do you pick?
In this scenario, my gut points strongly to killing the person who has already lost a spouse, but then I think I lean towards option B in the original question… does this change anyone else’s opinion?
@Scott H: “this evil Landsburg dictatorship”
Shhhhh! Don’t give him ideas.
While debating this is interesting it’s a bit like debating what the Rorschach ink blot really looks like.
@ Ken B
Agreed. This surrogate question project doesn’t seem to be taking us anywhere.
Hmmmm… well, on second thought, this is Landsburg’s evil dictatorship. He can decide if it’s taking us anywhere or not. =)
“nobody-really gets my vote. The preferences of the guy making the decision are what’s relevant.”
The question implies that you’re trying to maximize utility to the people involved. As such, you don’t actually have any preferences–how to maximize utility is not something you choose, it’s something you figure out.
“For goodness’s sake, they’ve told you what to do”
I’m inclined towards Option A for this reason. But the only justification for this unhappy outcome I can think of without involving the preferences of dead people (but trying to stay in the spirit of the problem) is that choosing Option A will signal that I respect peoples’ wills. If I choose Option B then no one will trust me to make these decisions in the future.
They have not told you what to do. They’ve told you what to do conditional on their death. They have not indicated their wishes should they survive. You don’t know the degree to which any of them wishes to remember their dead spouse.
which, btw is the point. I think Landsburg is saying, look, these people in the real world vote because they know what’s going to happen to them. In the analog of these scary situations, its as if people don’t know if they are the older or younger generation and they don’t know what they’ll have to live with.
I wonder if Steve is trying to recreate, virtually, Zimbardo’s famous study where he divided Stanford student subjects into guards and prisoners and asked the former to apply electric shocks to the latter. If so, I want to go on record as being morally opposed to killing these people.
A.
And I would hope that if anyone chose B the couple that lives would kill you for clearly violating their desire in a life and death situation.
Only an economist can go can against the wishes of ALL parties and still try to claim the moral high ground — I’m not saying that Steve is one of those economists, but I will say that this is where some of the behavioralists are going.
This is why no one should ever fund a foundation today, or be an organ donor. No one will respect your dying wishes.
Phil King: “They have not told you what to do. They’ve told you what to do conditional on their death. They have not indicated their wishes should they survive. You don’t know the degree to which any of them wishes to remember their dead spouse.”
I dispute this.
Steven Landsburg: “Therefore all four ask you, please, to choose A so that anyone who dies will be remembered by a loving spouse.”
If being remembered after death is important then Alzheimer’s risk becomes germane. Indeed all sorts of other things come into play. What is the likelihood that the surviving spouse is starving or being brutalized while doing the fondly remembering?
My contention is that there is no rule or Utilitarian calculus which yields predictable results such that they are also robust to small changes in the information set, or where hysteresis, concurrency or other processing problems don’t screw things up big time.
Assume each person makes $100,000 per year. If you kill one person from each couple “household income” will be cut in half. If you kill one couple, household income will remain unchanged. If you care about the “middle class” you will kill the couple. :)
Choice A because that’s what they want.
Martin
I understand that they all agreed on A, but they did it because they all want to be remembered. They’re displaying the thought process that they all think they’re not going to have to live with the consequence. If they said instead, A is best because I’d want to be remembered more than I wouldn’t want to remember you get closer to A.
Still, once you die you don’t get to impose your will anymore. You had your chance to make your mark, now it’s the next generation’s turn to mark the world. What about wills, you might ask? Until there’s a better way to split up an estate, they’re okay.
Dishonoring the will of dead for the benefit of the living is the best option. We suffer enough from prolonged effects of generations past whose wills carry through due to a societal sentiment and adherence to tradition. As a society we could do much better by breaking tradition, or at the very least questioning it. We should make all efforts to ensure older generations make a safe and pleasant exit but afterward there’s no need to live out their wishes.
Phil King: Thanks for your reply.
“If they said instead, A is best because I’d want to be remembered more than I wouldn’t want to remember you get closer to A.”
I think it’s reasonable to assume that these people are rational enough to have evaluated the risk of being the surviving spouse. This is a Landsburg problem after all. Have you seen his blue-eyed native puzzle?
“Dishonoring the will of dead for the benefit of the living is the best option”
I agree with your example of cultural tradition, but I disagree about wills. There are many people, young and old, who have preferences about what should happen after they die (at least I’m assuming there are such people). To dishonor the will of a dead person is to dishonor the preference of these living people as well, and that’s a cost I see no reason to ignore.
Example: Robbie wants his estate to go to the Dallas Cowboys once he’s dead. Tom wants his own estate to go to Pizza Hut once he’s dead. Tom dies tragically, but instead of Pizza Hut his estate goes to someone else. Tom’s preferences are no longer an issue, but now Robbie knows his estate won’t go to the Dallas Cowboys. It’s Robbie’s preference that is being ignored, and he is just as alive as you or me.
As Nelson Muntz might say:
Dishonoring a will is a victimless crime. Like punching someone in the dark.
Reputation matters. I would go so far as to say that any of these thought experiments that are supposed to assume that the rest of the world doesn’t know what you did are meaningless (see Muntz, 1995). Not to say that that’s how this question was originally framed. Just trying to clarify.
So a variant of the question might include:
(1) After you choose, you will be brought before a panel of 1000 people chosen at random, and they will decide whether you chose wisely. If they decide you didn’t, you be shot.
(2) After you choose, you will be brought before a panel of 1000 behavioral economists (or progressives, or, ….)
How would your choice change depending upon the panel selection? Why?
djp: “I would go so far as to say that any of these thought experiments that are supposed to assume that the rest of the world doesn’t know what you did are meaningless (see Muntz, 1995)”
With this point in mind I have a variant on your questions.
(1) “After you choose, you will be brought before a panel of 1000 people chosen at random, and they will decide whether you chose wisely. If they decide you didn’t, you be shot.”
After making their decision, each member of the panel will be brought before a new panel of 1,000 ordinary people (or behavioral economists, or…) who will decide whether he or she chose wisely. If the answer is no then I don’t like where this is going.
Martin,
That’s my point though. Sure Robbie’s preference is being ignored. But that’s ok! From my perspective, your preferences about what happens to you after you die can be violated. To me, Robbie et al haven’t come to terms with the fact that some day they will no longer have an impact and are trying to impose their preferences past expiration.
You get your time to express your preferences. That natural momentum of the past carries more than enough into the future. There’s no reason to tip the scales further.
Phil:
Does that mean it would be a lesser crime to kill someone than to maim them?
Maiming them affects their preferences before the event and after the event. Killing them does too. But after you’ve killed them, they’re dead. And you seem to be saying that after they’re dead their preferences no longer matter.
My initial thought was A – they have told you so do it. I then reflected a bit further. By doing what they said, I relieve myself of responsibility of the choice. This is not necessarily the best way to behave.
I find it doubtful that being remembered by a spouse would be such a mnotivating factor, I find it unlikely that anyone would think this so important. This makes me less inclined to respect their wishes.
What if each secretly hated their spouse and wanted to make sure they suffered through grief? They are then saying that they wanted to be missed, rather than remembered. Each could ask the same thing, but I would not feel the need to respect their wishes this time.
The one-offness is important. Also I assume that each would simply die without pain or knowledge of the outcome. Usually it is important to respect the wishes of the dead, because the living require it for smooth running of their affairs. If there is no effect on any other living people, then there is arguably no real need to consider the wishes of the dead.
So provided that the situation is unique, and the whole thing entirely secret, then probably B. The only people who knew would probably be glad.
Perhaps it would be better if you could save two people, rather than kill two? All four have been poisoned and are going to die, but you have enough antidote to save two. All slip into unconsciousness, and you have a bit of time to decide to which two to admminister the antidote. Is the answer different?
I choose A. For goodness’s sake, they’ve told me what to do.
a. They’ve told me. Presumably all 4 people are better placed to evaluate the post-killing situation since they would be doing the dying, surviving and remembering
Assumptions:
a. each couple has discussed it amongst themselves. I am not sure if it matters the two couples have discussed the matter with each other. Of course if one couple wants A and the other wants B, I would shoot the one which chose B (but that would not be random then, would it?)
b. Alan Wexelblat’s assumptions
c. No externalities imposed on anyone outside the 5 people involved (the 2 couples and the executioner)
As for Harold’s hypothetical I would still choose A (given everything else is the same).
Phil King:
“I understand that they all agreed on A, but they did it because they
all want to be remembered.”
No, that’s the motive they explicitly stated. That doesn’t mean
that being remembered is their sole motive for choosing A. If I say
I’d rather have a steak than your food bar because steak is better
for me, that may be a true motive. I may also want the steak
because it tastes better. If you’re so sure you know what’s better
for me, you may still try to foist your bar on me because it’s
vitamin and mineral laced.
People seldom choose something for one and only one reason.
Choosing B implies you’ll ignore people’s choices unless they
convince you otherwise. The usual optimum way to deal with someone
like that is to express a decision and refuse to state reasons.
“Just assume I have good reasons that I don’t wish to share.”
It is very difficult to think of an example of what is trying to be tested that seems real. One factor is should we care about the wishes of the dead after they have gone.
Say I come across a dying stranger. They have just enough time to ask me to do some small task after they have gone. Nobody else would have any knowledge of it – it is not passing on a message or suchlike.
I would prefer to perform the task, even though I believe they would have no knowledge either way. So I do respect to some extent the wishes of the dead. However, if it was a very onerous task, I would not do it, so I don’t respect their wishes all that much.
Why I have any respect at all for their wishes may not have a rational basis. I would suffer a small loss, and nobody apparently gains. It has something to do with empathy.
So back to our original example, we would respect the wishes of the “victims”, if there were no other considerations, because we have some consideration for the dead. However, we feel (rightly or wrongly) that there would be greater happiness if we choose the option that goes against their wishes. Whether we chose A or B depends which of these we value more.
@Harold: Bennet Haselton made, and I expanded on, a similar point, in a way that might make your impulse seem more rational. In a repeated game such impulses can direct us to the better choices; they are part of choosing the rational repeated stratgy
Why do we really obey the wishes of those who die? Like where to bury them for instance. I would say that we do it for the people who are still alive but has some connection to the one who died. And this is only because those people will otherwise feel screwed over somehow, even if it’s just because they know that their dead friend wanted something else. If you can find a person who has no friends or relatives, does it matter if we bury him according to his wishes or not? I would say no, because no one will ever feel like an injustice has taken place if we don’t. Or to put it differently, let’s say someone with no friends or relatives dies while swiming in the sea and the body is eaten by some fish but he really wanted to be creamated, who is negatively impacted by that?
According to what I just wrote we should choose B only if someone else knows and cares about their wish, which I guess is at least the other couple and potentially the guy pulling the trigger…
Oops, that should be option A in my previous post.
Ken B: I agree with what you said: “Lots of research shows that people implicitly apply repeated-game logic to one-time games. It’s almost impossible not to.”
To me this seems a non-rational strategy – it would surely be more rational to apply one-time logic to one-time games.
It becomes sort of logical if we must apply the same logic to all similar situations. Thus we may have a choice, either apply one- time logic to all situations, or apply repeat-game logic to all situations. If we mostly meet repeat-game situations, then it makes sense to apply repeat-game logic to all situations. But this means we have imposed a rule that prevents the best outcome in every situation.
We have developed systems which work best in the majority of situations, and we cannot change these in unusual situations. Our rule says “respect the wishes of the dead”. This is usually a good rule, because it helps the living, includimng ourselves. In the occasional situation where respecting the wishes of the dead may not be the best thing to do, we cannot prevent ourselves from applying the rule anyway.
@Harold:”To me this seems a non-rational strategy – it would surely be more rational to apply one-time logic to one-time games. ”
That is really the source of most of the embarassment in experimental GT. People just don’t.
There’s an inherent difficulty. How exactly do you study the one time game? You have to convince people it is one time, and then people not used to one time games need to learn to play them. you need to retrain their instincts, which are for repeated games. How do you learn to play a game? By playing it repeatedly …
So here we have another example of consistent and predictable non-rational behaviour. It gives scope for policy makers to go against the wishes of the people for their own good. Examples may be events only ecountered by the individual rarely, where repeated-game logic may lead one astray.
Option B is the most rational choice. Once you’re dead, it doesn’t mattered how you’re thought of. It’s human nature to try to use being well remembered as a way to attain a sort of life after death. But ultimately it doesn’t matter…you’re dead.
Killing one person from each couple would effectively destroy four lives. Two people would be dead, and two others would face years of loneliness or depression.
(though the question didn’t stipulate that the couples were happily married. it is possible that the death of one spouse would be welcomed by the other, which might explain the calls for “please, only kill one of us [the other one]”)