I.

People love trying to find holes in the drowning child thought experiment. This is natural: it’s obvious you should save the child in the scenario, but much less obvious that you should give lots of charity to poor people (as it seems to imply). So there must be some distinction between the two scenarios. But most people’s cursory and uninspired attempts to find these fail.

For example, some people say the difference is distance; you’re close to the drowning child, but far from people dying in Africa. Here are some thought experiments that challenge that:

You’re a surgeon, using a telepresence robot to operate on someone in China. In the middle of the operation, a medical student watching in the Chinese side starts choking. He is the only other person in the room (besides your patient, who is unconscious), so nobody else can help him. Your robot can do the Heimlich Maneuver and save their life, but it would delay the surgery five minutes while you readjusted the settings afterwards, which would make you late for lunch. Should you save them?

Here it seems obvious that you should save them, even though they’re all the way in China.

Is the problem that “you” are sort of “in” China via your robot, even if not physically? Here’s another example:

The Dublin-NYC Portal is an art installation with branches in Dublin and NYC. If you stand in front of one of the portals, you can see and (not in real life, but in our thought experiment) hear the people on the other side. It’s usually too crowded to really interact with, so one day you (in NYC) go to the installation at 3 AM and are the only person around. You start talking to two Dubliners, who are the only people around on their side of the portal. Suddenly, one of them starts choking. The other shrieks “I need to do the Heimlich Maneuver, but I don’t know how!” You know how, and could easily walk them through it. But it’s cold and you want to go home. Should you help them?

Again, the answer is clear even though you’re 3000 miles away. At this point, saying that you’re “virtually” in Dublin seems like a stretch. Here the issue seems to be some sort of entanglement. But it’s hard to say exactly how the entanglement works, and it doesn’t seem to be a simple one-to-one correspondence where you’re the only person who can help. For example:

You’re at the Sociopathic Jerks Convention (you’re neither sociopathic nor a jerk - you’re the caterer). Everyone is on the lawn of the conference center, waiting for one of the sessions to begin, when you all notice a child drowning in the lake nearby. Along with yourself, there are 1,000 sociopathic jerks. But at the last session, someone took a poll on exactly this question and everyone agreed they wouldn’t lift a finger to help; either you save the child, or nobody does. Do you jump in and save them?

Here it seems like the sociopathic jerks might as well be furniture - their presence doesn’t change your situation compared to the scenario where you’re there alone.

II.

TracingWoodgrains draws off a now-deleted essay by Jaibot which talks about the “Copenhagen interpretation of ethics”. It argues that by “touching” a situation - a vague term having something to do with causal entanglement - you gain moral obligation for it. If you can simply avoid touching it, your moral obligation goes away.

I think this explains half the problem, but I can think of another half that it doesn’t explain. Consider:

You inherit a beautiful cabin in the woods. It’s downstream of a vast semi-magical megacity which is a little denser than should be physically possible. Every time a child falls into any of the megacity’s streams, lakes, or rivers, they get swept away and flow past your cabin; there’s a new drowning child every hour or so.

Assume that all unmentioned details are resolved in whatever way makes the thought experiment most unsettling - so for example, maybe the megacity inhabitants are well-intentioned, but haven’t hired their own lifeguards because their city is so vast that this is only #999 on their list of causes of death and nobody’s gotten around to it yet.

Here I’m split on whether the Copenhagen hypothesis works. A person who lives in the cabin and fails to rescue every child seems much less monstrous than someone who only ever encounters the situation once, even though both of them “touch” the situation exactly as much. Still, as the hypothesis predicts, we are less comfortable with this situation than the normal one where you live far away from the cabin and never worry about it - living near the cabin (“touching” the situation) seems to have some moral impact.

Here’s somewhere I think Copenhagen more clearly fails:

Your regular house burns down, and you have no choice but to live in the cabin for a while. You think to yourself: “Well, every day, I’ll rescue one child, then not worry about it. This is better for the children, since it increases their survival rate from 0 to 1/24. And it’s better for me, since I’m not homeless while I wait for them to rebuild my house). It’s a win-win situation.

So you do this for a few months, and then they rebuild your house, and you move back to your regular hometown. You live there for five years without incident. One day, you see a child drowning in the local river. You don’t recognize them, so it can’t be the child of any of your fellow citizens - probably their parents are some of the travelers who pass through this town on the way to greener pastures. You are late for an important business meeting, you’re wearing a nice suit, and it would be especially annoying to jump in a river right now. In fact, you estimate that it is 10x costlier to rescue this child than any of the children who you encountered at the cabin so long ago. What do you do?

Here Copenhagen fails to predict a difference between refusing to rescue the 37th kid going past the cabin, vs. refusing to rescue the single kid in your hometown; you are “touching” both equally. But I think most people would consider it common sense that refusing to rescue the 37th kid near the cabin is a minor/excusable sin, but refusing to rescue the one kid in your hometown is inexcusable.

Again sticking to a purely descriptive account of intuitions, I think this represents a sort of declining marginal utility of moral goods. The first time you rescue a kid, you get lots of personal benefits (feeling good about yourself, being regarded as a hero, etc). By the 37th time, these benefits are played out. If you refuse to rescue a child for the relatively high benefits of a single situation, we think you must have no moral sense at all. But if you fail to rescue them the 37th time, we think this is pretty understandable and similar to what we would do in the same situation.

(This “declining marginal utility” explanation is less natural than something like “the obligation to rescue all those children is ruining my life”. But I think it’s more accurate; if we come up with a thought experiment where it doesn’t ruin your life in any way - where it only takes a few hours from your day and you have enough left to accomplish everything you need - then it still seems harsh to demand someone rescue 37 children every day. And when there is an actual moral obligation - like parenting your own children - we don’t accept “it will ruin my life” as an excuse to get out of it.)

III.

So these two descriptive theories - the Copenhagen hypothesis, and the declining marginal utility of moral goods - do a good job explaining our intuitions.

But some people leap from there to saying they’re also the right prescriptive theories - they determine what morality really is, and what rules we should follow.

I think this is a gigantic error, the worst thing you could possibly do in this situation. These are essentially rules for looking good to other people. To follow them is to say that you will always optimize for seeming cool, no matter how many people you have to kill in order to do it. So for example:

Because of Copenhagen hypothesis, you decide it’s morally wrong to go to the cabin - by “touching” the situation, you would be accepting blame for all of the drowning children.

But you really want a vacation away from your polluted city, and you can’t afford any other vacation home, so you’re always looking for some way to solve the problem. One day, you encounter a man with a giant truck. He says that for $525, he could uproot your cabin, load it on his truck, and deposit it on unoccupied land near a small camp a few miles away. In this new location, you couldn’t hear the children screaming. In fact, there are some other cabins at the camp, none of them ever think about the death river, and nobody ever blames them for it.

You’re pretty excited about this. But you hear that your wife’s sister’s friend’s niece has a rare disease that would cost exactly $525 to cure (and she is poor, and can’t afford it). Normally you wouldn’t care much about someone so distant from you. But your wife asks - hey, isn’t it sort of hypocritical to move the cabin “in order to be a moral person”, when you’re in fact not helping anyone in any way? Wouldn’t it actually make you more of a moral person to spend the $525 curing her sister’s friend’s niece, then vacation at the cabin using earplugs to not hear the screaming kids (which wouldn’t result in any more kids dying than not going to the cabin, or moving the cabin)?

This is all awkward enough that maybe you want to push the Copenhagenness back a step and just refuse to touch the cabin at all. Refuse to inherit it, lock your door, tell the lawyer who says you own it now that he needs to get off your property or else you’ll shoot. But we can still make your life difficult:

You live in a house in the suburbs. You never even considered living in a cabin in the woods. However, the nearby semi-magical megacity plans to build a dam. This would rearrange its various internal waterways so that all the drowning children would get carried by the current through your backyard; you would be in the exact same situation as the cabin owner. All of your life savings are tied up in your mortgage, the mere threat of this dam has crashed the value of your house, and you can’t afford to move. Luckily, a lobbyist owes you a favor. She offers to repay you in one of two ways.

First, she could lobby the megacity to redirect the dam; this would cause the drowning children to go somewhere else - they would be equally dead, but it’s not your problem.

Second, she could lobby the megacity to hire a part-time lifeguard (full-time is beyond her power), who could save half the drowning children. This still ruins your life (a drowning child passes your house once every two hours), and still crashes your home value to zero, but thousands of children would be saved per year. Also, she has a side business as a handyman, and could install double-pane glass windows on your house so you couldn’t hear the children screaming.

The Copenhagen theorist would be in a bind here. You really want to avoid the dam forcing you to “touch” the situation, and then either spend your whole life saving children, or be culpable for failing to do so. But it seems both heartless and pointless to waste your one lobbyist favor on a river redirection which doesn’t change anything about the real world (as many children will die as ever) when you could instead use it to do lots of good.

My best bet for how a thoughtful Copenhagener would respond is that they would say you had terrible moral luck by happening to end up where the dam was going to redirect the drowning children; however, this itself caused you to “touch on” the situation and now you can be judged for how you respond (including your cowardly response of trying to redirect the river somewhere else).

I don’t buy it.

Unfortunately, the lobbyist dies of a heart attack before you can call in either favor. The dam is built and the children pass by your house. You save one per day, but not all of them.

You have a neighbor. The neighbor lives far enough away that the river doesn’t affect him at all; he couldn’t hear the children in any case.

When you tell him this story, he calls you a monster - how could you only help one child per day?

You agree this is inadequate, and decide to hire a lifeguard yourself to solve the problem. You offer to split the costs with your neighbor: you will pay 80%, he’ll pay 20%.

The neighbor says haha, no way, he’s not going to waste 20% of his money helping some kids he has no relationship with. But he also thinks you’re a monster unless you pay the money.

God notices there is one extra spot in Heaven, and plans to give it to either you or your neighbor. He knows that you jump in the water and save one child per day, and your neighbor (if he were in the same situation) would save zero. He also knows that you are willing to pay 80% of the cost of a lifeguard, and your neighbor (who makes exactly the same amount of money as you) would pay 0%. However, in reality, the river and the drowning children are going by your house, not your neighbor’s house. Which of you should get the spot in Heaven?

Here it seems obvious that you are a better person than your neighbor. But then what remains of the “moral luck” explanation? What remains of Copenhagen, where you are blamed for a situation if you touch it?

Maybe you have to choose to touch it for it to count? But this seems false; in Singer’s original drowning child experiment, you didn’t choose to be the only person near the lake when the kid was drowning. It was just a weird coincidence.

In fact, it seems like we all benefit from the same sort of moral luck as the neighbor. Suppose Alice is born in a gated community in the US, to a family making $200,000/year; she goes to her local college, stays in her rich hometown, and eventually makes $200,000/year herself. There are no poor people near her, so she has few moral obligations. But Bob is born in Zimbabwe, to a rare upper-class well-connected Zimbabwean family making $200,000 year; he inherits his father’s business and also makes $200,000 year himself. But he lives in the middle of horrible poverty. His housekeeper is dying of some easily-cured disease, all of his school friends are dying of easily-cured diseases, every day when he goes to work he has to walk over half-dead people screaming for help. It seems like Alice got lucky by not being Bob; she has no moral obligations, whereas he has many. Suppose that Bob only helps a little bit, enough that we would consider him pretty stingy given his situation - maybe he helps his absolute closest school friend, but lets several other school friends die. And suppose that if Alice was in Bob’s situation, she would do even less, but in fact in real life she satisfies all of her (zero) moral obligations. If there’s only one spot in Heaven, should it go to Alice or Bob?

Someone who’s still desperately trying to preserve Copenhagen would have to say that the “one spot in Heaven” prompt isn’t fair - God presumably has His own criteria which exploit His perfect omniscience, but we humans must think about morality on a merely human level. I still don’t buy it. For one thing, God isn’t using any special omniscient knowledge that we (the people reading this thought experiment) don’t also have and use easily. For another, if you’re even slightly religious, actually getting the literal spot in Heaven should be one of the top things on your mind when you’re deciding whether to be moral or not. Even if you’re atheist, trying to be the sort of person who would get a spot in Heaven, if it existed, seems like a worthier goal than whatever the Copenhagen-follower is doing.

IV.

So again, the question is - what is the right prescriptive theory that doesn’t just explain moral behavior, but would let us feel dignified and non-idiotic if we followed it?

My favorite heuristic for thinking about this is John Rawls’ “original position” - if we were all pre-incarnation angelic intelligences, knowing we would go to Earth and become humans but ignorant of which human we would become, what deals would we strike with each other to make our time on Earth as pleasant as possible? So for example, we would probably agree not to commit rape, because we wouldn’t know if we would be the offender or the victim, and we would expect rape to hurt the victim more than it helped the offender.

Here we would probably agree to save drowning children, because if we were involved in the situation at all, we would have a 50% chance of being the rescuer (minor inconvenience) or the child themselves (life or death importance).

But we would also agree to save people dying of easily-cured diseases in the Third World, because we wouldn’t know if we would be those people either. Everyone would agree to a proposed deal that rich people donate a small fraction of their income to charity, because it would be only a mild inconvenience if they turned out to be rich, but a life-saver if they turned out to be poor.

Further, since we wouldn’t know whether we would be Alice (low level of moral obligation) or Bob (very high level of moral obligation), we would take out insurance by agreeing that everyone needed to pay the same modest amount into the general pot for helping people.

(How much should they pay? Enough to pick the low-hanging fruit and make it so nobody is desperately poor, but not enough to make global capitalism collapse. I think the angelic intelligences would also consider that rich people could defect on the deal after being born, and so try to make the yoke as light as possible.)

A final deal might look like this: we’ll all cooperate by sending a bit of our money to a general pot for helping people in terrible situations. And if there’s a more urgent situation that group contributions can’t help - because for example a child is drowning right now and there’s only one person close enough to save them - then we’ll deputize that one person to save them, and assume it will all even out in the end.

(Actually, even better would be pay that person a reward for their trouble out of the general pot - then there’s no unfairness or special obligation on one person rather than another!)

Here we’re able to bring back all of those things we rejected earlier - proximity, urgency, being the only person available - not because they determine who is worthy of being saved, but in the context of a coalition that plans to save everybody but which in an emergency needs to act through whoever is available. This is no different from a police force which, learning of a serious crime in progress, asks the officer closest to the site to respond, even if that officer isn’t a specialist in that particular type of crime, or even if that officer is one minute away from clocking out and it’s unfair to make them work overtime.

All of this makes perfect sense - except that the coalition is in arrears, there is no general pot, and most bad things go unprevented. Only the extra “save people close to you” rule, tacked on as an afterthought, still functions, because that one makes people look good when they do it and is easier to enforce through reputational mechanisms.

I think you should probably still save someone close to you (eg drowning), partly because this rule is valuable even on its own (ie it’s better to do it than not do it), and partly because, since other people are following it, you actually have a reciprocal obligation to your fellow coalition members here (ie you expect that if your child was drowning, someone else would help, so you’re free-riding if you don’t help them).

If you end up at the death cabin, you don’t have an obligation to save every single child who passes by, because the coalition didn’t intend for the “save drowning children” obligation to be an unusual burden on anyone in particular, and because nobody else is doing this so you’re not betraying fellow coalition members. People may incorrectly think less of you if you don’t do this, and you might want to take action to avoid reputational damage, but this isn’t a moral obligation. The real answer to this problem is that the coalition should split the cost of hiring a lifeguard - or, if for some reason you are the only person who can be in the area, compensate you for your time. Given that the coalition isn’t strong enough to actually do these things, your obligations are limited, and not made any better or worse by living in the cabin vs. further away.

I think it’s virtuous, but not obligatory, to behave as if the coalition is still intact, and try to give a portion of your income to some sort of virtual version of the general pot. You could also think of the government as some sort of very distorted flawed real-life version of the coalition and consider your obligations fulfilled by paying taxes, but I think this is an insult to the angelic intelligences, and you should just go with whatever seems like the closest thing to their original plan without waiting for it to actually be instantiated.

I think this is more dignified than the thing where you try to hire someone for $525 to move your cabin to a different location so you don’t feel like you’re “touching” the problem, or whatever.