Because it seems incredibly unlikely to maximize utility,
Avoiding for the moment the question whether utilitarianism is the right approach to these kinds of problems. There is in fact a decision theory argument in favor of this. Eliezer stumbled up on a version of it and didn’t react well, specifically banning all detailed discussion of it from LW in an extremely ham-handed manner.
neither does it accord with what seems to me a general principle that punishment should be at most proportionate to the crime being punished.
Where does this principal come from? Can you provide any utilitarian justification for it? It’s a useful mildly useful Schelling point in certain rather specific circumstances but that’s about it.
There is in fact a decision theory argument in favor of this. Eliezer stumbled upon a version of it [...]
I haven’t given a lot of thought to that particular argument, but it doesn’t look to me as if it can plausibly work here. In order for it to produce useful effects without unacceptable collateral damage, it seems to me that you want the threat known to a small number of people and to persuade them to work towards a highly specific goal that those people are particularly well-suited to achieving. Exactly how small, how specific, what goal, and how well-suited may vary together in kinda-obvious ways. These conditions don’t seem to me to be anywhere near to being met in the case of the Christian idea of hell.
Can you provide any utilitarian justification for it?
I’m not sure. If I’m thinking in utilitarian terms (which I generally am) I am mostly concerned with actions rather than principles (unless the question at issue is “would this be a good principle to have?”, which it isn’t here). I brought the deontological viewpoint in at all only because I know some people reject consequentialism and may be (or profess to be) wholly unmoved by the mere fact that something causes vast amounts of unnecessary suffering.
it seems to me that you want the threat known to a small number of people and to persuade them to work towards a highly specific goal that those people are particularly well-suited to achieving.
Not really. In fact one reason for universality is to discourage reactions like Eliezer’s.
The more people the threat is known to, the less likely that they all comply. Then, if the idea is to actually follow through on the threat (note that if that isn’t the idea then you don’t have here any sort of argument that hell is not morally monstrous) you’re stuck with torturing a bunch of people for ever, which is liable to outweigh the benefits.
The more people the threat is known to, the less likely that they all comply.
And someone who doesn’t know about it is even less likely to comply. If you’ve already concluded that threatening to torture someone is worth it for the increased chance of getting compliance, then the exact same calculation applies to everyone else.
Hmm, that’s true. So let’s see how that calculation works out in the two cases being considered.
The one that upset Eliezer: the goal is to get a sufficient number of people from a rather unusual population to do something extraordinary. Probably doesn’t actually work (see e.g. the actual responses to that LW post, which AFAIK didn’t in any case take the form “OK, I must start doing all I can to comply”) but the idea here is that a small number of people can make a huge difference. The goal, therefore, is a very large impact per expected person tortured.
The one we’re talking about here: well, we can see roughly how this turns out. Empirically, people who believe in the Christian hell don’t behave dramatically better than people who do. Hence, a fairly small impact per expected person tortured.
(If I try to steelman your argument, I get things like this: “Ah, but you’re looking now, at a time when in fact scarcely anyone actually believes in hell. Historically, most of the hell-belief comes centuries ago when Christianity was stronger, and for that reason it was effective then, which was an essential element in making Western mostly-Christian civilization the effective, high-trust thing it is; so we may suppose that God gave people the idea of hell with that purpose, knowing that it would die out once it was no longer needed.” I don’t want to get into a lengthy response to an argument that I just made up and that you may think is no good, but here’s a sketchy one: 1. Plenty of people still believe in hell. 2. The doctrine of hell whose (de)merits we’re discussing doesn’t actually say that people are only eligible for hell if they have never stopped believing in it. 3. It’s not in fact at all clear that the doctrine of hell was ever very effective in improving behaviour. 4. Neither is it at all clear that any such improved behaviour had much to do with the evolution of Western society into what it is now. 5. It’s not hard to think of other ideas a benevolent superbeing could have given to its chosen people which would in fact have done more good.)
Empirically, people who believe in the Christian hell don’t behave dramatically better than people who do.
Hasn’t quite been my experience but, whatever.
The doctrine of hell whose (de)merits we’re discussing doesn’t actually say that people are only eligible for hell if they have never stopped believing in it.
Of course, otherwise it would be completely useless as it would simply motivate people to stop believing in it.
I agree that hell is a bad idea. That said, many Christians are tending more toward the position that very few people actually go to hell. It could be that only the very worst people go there, together with a few of the better people who could have done far more good than they actually did. If the numbers are small enough, there might be a significantly larger number of people who did a great amount of good, but would not have done it, if they had not believed in hell. In that case, there might be a large impact per person going to hell.
Obviously, there are still many problems with this, like whether that large impact can justify anything eternal.
Avoiding for the moment the question whether utilitarianism is the right approach to these kinds of problems. There is in fact a decision theory argument in favor of this. Eliezer stumbled up on a version of it and didn’t react well, specifically banning all detailed discussion of it from LW in an extremely ham-handed manner.
Where does this principal come from? Can you provide any utilitarian justification for it? It’s a useful mildly useful Schelling point in certain rather specific circumstances but that’s about it.
I haven’t given a lot of thought to that particular argument, but it doesn’t look to me as if it can plausibly work here. In order for it to produce useful effects without unacceptable collateral damage, it seems to me that you want the threat known to a small number of people and to persuade them to work towards a highly specific goal that those people are particularly well-suited to achieving. Exactly how small, how specific, what goal, and how well-suited may vary together in kinda-obvious ways. These conditions don’t seem to me to be anywhere near to being met in the case of the Christian idea of hell.
I’m not sure. If I’m thinking in utilitarian terms (which I generally am) I am mostly concerned with actions rather than principles (unless the question at issue is “would this be a good principle to have?”, which it isn’t here). I brought the deontological viewpoint in at all only because I know some people reject consequentialism and may be (or profess to be) wholly unmoved by the mere fact that something causes vast amounts of unnecessary suffering.
Not really. In fact one reason for universality is to discourage reactions like Eliezer’s.
The more people the threat is known to, the less likely that they all comply. Then, if the idea is to actually follow through on the threat (note that if that isn’t the idea then you don’t have here any sort of argument that hell is not morally monstrous) you’re stuck with torturing a bunch of people for ever, which is liable to outweigh the benefits.
And someone who doesn’t know about it is even less likely to comply. If you’ve already concluded that threatening to torture someone is worth it for the increased chance of getting compliance, then the exact same calculation applies to everyone else.
Hmm, that’s true. So let’s see how that calculation works out in the two cases being considered.
The one that upset Eliezer: the goal is to get a sufficient number of people from a rather unusual population to do something extraordinary. Probably doesn’t actually work (see e.g. the actual responses to that LW post, which AFAIK didn’t in any case take the form “OK, I must start doing all I can to comply”) but the idea here is that a small number of people can make a huge difference. The goal, therefore, is a very large impact per expected person tortured.
The one we’re talking about here: well, we can see roughly how this turns out. Empirically, people who believe in the Christian hell don’t behave dramatically better than people who do. Hence, a fairly small impact per expected person tortured.
(If I try to steelman your argument, I get things like this: “Ah, but you’re looking now, at a time when in fact scarcely anyone actually believes in hell. Historically, most of the hell-belief comes centuries ago when Christianity was stronger, and for that reason it was effective then, which was an essential element in making Western mostly-Christian civilization the effective, high-trust thing it is; so we may suppose that God gave people the idea of hell with that purpose, knowing that it would die out once it was no longer needed.” I don’t want to get into a lengthy response to an argument that I just made up and that you may think is no good, but here’s a sketchy one: 1. Plenty of people still believe in hell. 2. The doctrine of hell whose (de)merits we’re discussing doesn’t actually say that people are only eligible for hell if they have never stopped believing in it. 3. It’s not in fact at all clear that the doctrine of hell was ever very effective in improving behaviour. 4. Neither is it at all clear that any such improved behaviour had much to do with the evolution of Western society into what it is now. 5. It’s not hard to think of other ideas a benevolent superbeing could have given to its chosen people which would in fact have done more good.)
Hasn’t quite been my experience but, whatever.
Of course, otherwise it would be completely useless as it would simply motivate people to stop believing in it.
I agree that hell is a bad idea. That said, many Christians are tending more toward the position that very few people actually go to hell. It could be that only the very worst people go there, together with a few of the better people who could have done far more good than they actually did. If the numbers are small enough, there might be a significantly larger number of people who did a great amount of good, but would not have done it, if they had not believed in hell. In that case, there might be a large impact per person going to hell.
Obviously, there are still many problems with this, like whether that large impact can justify anything eternal.
How so?