Even if we’re willing to take it out of context like this, we might still consider it ethically undesirable to have kids in a time and place where military conflict or politically caused poverty is likely.
I personally wouldn’t decide to have kids in a warzone… But what context are you referring to? Is there any context outside of sudden, subjectively unlikely disaster where the quote is meaningful?
I personally wouldn’t decide to have kids in a warzone…
...but it’s okay if others do it? How is that different from saying, “I personally woudn’t decide to abuse children...”
Is there any context outside of sudden, subjectively unlikely disaster where the quote is meaningful?
It was written by Michael Jackson. I don’t think he was referring to sudden, subjectively unlikely disasters, but the personal material means of people deciding to become parents.
How is that different from saying, “I personally woudn’t decide to abuse children...”
It’s impossible to have children and do no actions whatsoever which are less than optimal for the children. Rather, people make—and have to make—tradeoffs between things being bad for the children and other considerations. There is an acceptable range of such tradeoffs. Having kids in a warzone falls in that range and abusing kids does not. And even if you think people making other tradeoffs are actually wrong rather than just making the tradeoffs based on different circumstances, there are degrees of being wrong and abuse is wrong to a greater degree.
The dominating distinction between our perspectives is that I don’t think having kids in a warzone is an acceptable tradeoff, where you think it is.
This is probably just an intuitive disagreement about the relative harm and benefit of being born into a warzone.
I think it is clearly a very bad deal for the child, and to do it recklessly or out of selfishness in fact constitutes a form of child abuse. Of course, if you would actually rather be born into poverty or war, than not be born, you will disagree where the acceptable range lies.
We do not disagree about the rest of the argument.
It is possible that this would have been better overall.
Even if we reject 1, humanity was no where near extinction for thousands of years now.
You can easily augment the underlying harm avoidance principle with a condition that it should not result in the extinction of intelligent life (assuming that intelligent life doesn’t cause even more harm in the long run).
It’s a hypothetical—there is no evidence for or against it as it never happened and is highly unlikely to happen.
AI’s never been developed before either, but that hasn’t stopped people from trying to forecast the future, or at least enumerate likely scenarios. You were making a claim about a causal link: “following HedonicTreader’s guideline will cause humanity to go extinct.” Either this link exists or it does not, and you don’t get to back out of providing evidence just because the situation is a hypothetical.
But let me point out that it sets up a downward feedback loop.
What’s the loop?
(EDIT: Upon reflection, I feel I should clarify that I’m not actually disagreeing with you here. I’m slightly more sympathetic to HedonicTreader’s position than yours at the moment, but with some convincing, that could easily change. My questions should be interpreted more as requests for information than as rhetorical challenges.)
you don’t get to back out of providing evidence just because the situation is a hypothetical.
I use the word “evidence” to mean empirical evidence, that is, evidence from reality. Such does not exist in this case. Arguments from analogy, logic, etc. are not evidence.
What’s the loop?
The worsening of conditions triggers a major contraction of population which worsens the conditions further (contemporary economies take contraction badly and at sufficiently low population numbers and densities advanced technology becomes problematic) which triggers further contraction of the population...
His ethical guideline has nothing to do with how close humanity is to extinction.
Except I already wrote:
You can easily augment the underlying harm avoidance principle with a condition that it should not result in the extinction of intelligent life (assuming that intelligent life doesn’t cause even more harm in the long run).
You don’t even have to apply the principle of charity, you could just look at what I had literally written.
However, if practiced diligently, it can bring humanity to extinction in a few generations from any population size.
Nonsense. Most humans don’t live in a warzone at any time now. And followed in extreme poverty, this principle would reduce local malthusian traps and probably reduce poverty; at least the suffering of children from poverty.
Af first I thought you doing that Redditesque sarcasm, in which you argue a straw man of the outgroup in a mocking way, which made me disappointed since the goal is signaling rather than discourse.
However perhaps you are being serious? Are social services a valid means of feeding a baby, rendering the original quote not applicable in countries where social services exist? I think the answer is obviously yes, in that if social services are available, people are going to use them. Whether the should exist is a separate discussion.
I think it’s a law that if you fund something you get more of it. Serious discussion of safety nets, etc. already assumes some parasite response from the “ecosystem,” takes it into account, and argues safety nets are still a good thing on net.
I think “unintended consequences” is a better analysis framework than “parasite response from the ecosystem”.
And speaking of, there is a recent paper discussed on MR which claims to show how safety nets drive down the decline in labor force participation and, in particular, that “the Clinton-era welfare reforms lowered the incentive to work”.
I think “unintended consequences” is a better analysis framework than “parasite response from the ecosystem”.
It certainly sounds less cynical, unless we use strong charity and see it in the most technical way possible.
I think the most plausible use case for government-funded incentives to have extra kids is a wide consensus that a society doesn’t have enough of them at the time, according to some economical or social optimum.
But even this requires a level of cynicism in seeing kids as a means to an end.
I was being serious. Abstractly, if my doing X requires Y, but I don’t have Y but I’m confident that if I do X the government will give me Y, then my lack of Y isn’t much of a reason to forgo X.
Interesting… can you say more about why you include a term in that equation for internal negative value (what you label “suffering” here), but not internal positive value (e.g., “pleasure” or “happiness” or “joy” or “Fun” or whatever label we want to use)?
I suppose it was because the original quote started with a negative framing, the assumption that the baby might not be fed.
I think both birth and death are stressful experiences that are not worth going through unless there are compensating other factors. I don’t think infants have enough of those if they die before they grow up.
Also I suspect human life is generally overrated, and the positives of life are often used as an excuse to justify the suffering of others. I do not trust people to make a realistic estimate and act with genuine benevolence.
If you can’t feed your baby, then don’t have a baby.
-Michael Jackson (Wanna be starting something)
Right, just the thing they should have told those irrational pregnant women who ran away from the Eastern part of Ukraine.
Even if we’re willing to take it out of context like this, we might still consider it ethically undesirable to have kids in a time and place where military conflict or politically caused poverty is likely.
I personally wouldn’t decide to have kids in a warzone… But what context are you referring to? Is there any context outside of sudden, subjectively unlikely disaster where the quote is meaningful?
...but it’s okay if others do it? How is that different from saying, “I personally woudn’t decide to abuse children...”
It was written by Michael Jackson. I don’t think he was referring to sudden, subjectively unlikely disasters, but the personal material means of people deciding to become parents.
It’s impossible to have children and do no actions whatsoever which are less than optimal for the children. Rather, people make—and have to make—tradeoffs between things being bad for the children and other considerations. There is an acceptable range of such tradeoffs. Having kids in a warzone falls in that range and abusing kids does not. And even if you think people making other tradeoffs are actually wrong rather than just making the tradeoffs based on different circumstances, there are degrees of being wrong and abuse is wrong to a greater degree.
The dominating distinction between our perspectives is that I don’t think having kids in a warzone is an acceptable tradeoff, where you think it is.
This is probably just an intuitive disagreement about the relative harm and benefit of being born into a warzone.
I think it is clearly a very bad deal for the child, and to do it recklessly or out of selfishness in fact constitutes a form of child abuse. Of course, if you would actually rather be born into poverty or war, than not be born, you will disagree where the acceptable range lies.
We do not disagree about the rest of the argument.
Is it the same to die without ever abusing children and to die childless?
Applying this, humanity would have quietly died out a few thousand years ago...
2 responses:
It is possible that this would have been better overall.
Even if we reject 1, humanity was no where near extinction for thousands of years now.
You can easily augment the underlying harm avoidance principle with a condition that it should not result in the extinction of intelligent life (assuming that intelligent life doesn’t cause even more harm in the long run).
That makes no sense to me at all.
Because it did not follow the ethical guideline that you suggested.
This statement, if true, only shows that not following the guideline back then was the correct choice. What about today?
What, do you feel, is the relevant difference between back then and today?
It’s as HedonicTreader said: humanity is nowhere near extinction today.
His ethical guideline has nothing to do with how close humanity is to extinction.
However, if practiced diligently, it can bring humanity to extinction in a few generations from any population size.
This is a questionable claim. Do you have any evidence to support it?
It’s a hypothetical—there is no evidence for or against it as it never happened and is highly unlikely to happen.
But let me point out that it sets up a downward feedback loop.
AI’s never been developed before either, but that hasn’t stopped people from trying to forecast the future, or at least enumerate likely scenarios. You were making a claim about a causal link: “following HedonicTreader’s guideline will cause humanity to go extinct.” Either this link exists or it does not, and you don’t get to back out of providing evidence just because the situation is a hypothetical.
What’s the loop?
(EDIT: Upon reflection, I feel I should clarify that I’m not actually disagreeing with you here. I’m slightly more sympathetic to HedonicTreader’s position than yours at the moment, but with some convincing, that could easily change. My questions should be interpreted more as requests for information than as rhetorical challenges.)
I use the word “evidence” to mean empirical evidence, that is, evidence from reality. Such does not exist in this case. Arguments from analogy, logic, etc. are not evidence.
The worsening of conditions triggers a major contraction of population which worsens the conditions further (contemporary economies take contraction badly and at sufficiently low population numbers and densities advanced technology becomes problematic) which triggers further contraction of the population...
Except I already wrote:
You don’t even have to apply the principle of charity, you could just look at what I had literally written.
Nonsense. Most humans don’t live in a warzone at any time now. And followed in extreme poverty, this principle would reduce local malthusian traps and probably reduce poverty; at least the suffering of children from poverty.
I don’t think this is how the real world works.
Well, quite. What can we do about that?
What exactly is the problem you want to solve? :-/
Clearly.
But what if I can get taxpayers to feed my baby?
Af first I thought you doing that Redditesque sarcasm, in which you argue a straw man of the outgroup in a mocking way, which made me disappointed since the goal is signaling rather than discourse.
However perhaps you are being serious? Are social services a valid means of feeding a baby, rendering the original quote not applicable in countries where social services exist? I think the answer is obviously yes, in that if social services are available, people are going to use them. Whether the should exist is a separate discussion.
I think it’s a law that if you fund something you get more of it. Serious discussion of safety nets, etc. already assumes some parasite response from the “ecosystem,” takes it into account, and argues safety nets are still a good thing on net.
I think “unintended consequences” is a better analysis framework than “parasite response from the ecosystem”.
And speaking of, there is a recent paper discussed on MR which claims to show how safety nets drive down the decline in labor force participation and, in particular, that “the Clinton-era welfare reforms lowered the incentive to work”.
It certainly sounds less cynical, unless we use strong charity and see it in the most technical way possible.
I think the most plausible use case for government-funded incentives to have extra kids is a wide consensus that a society doesn’t have enough of them at the time, according to some economical or social optimum.
But even this requires a level of cynicism in seeing kids as a means to an end.
I was being serious. Abstractly, if my doing X requires Y, but I don’t have Y but I’m confident that if I do X the government will give me Y, then my lack of Y isn’t much of a reason to forgo X.
Can you also get them to pay for cryonics? I don’t know if you consider cryonics worthwhile, but the point is that “feed” generalizes easily.
Urahara Kisuke
The difference is that babies suffer if they starve, but not if they don’t have cryonics.
The badness of making an extra life comes from its suffering (+ negative externalities) [- positive externalities]
Interesting… can you say more about why you include a term in that equation for internal negative value (what you label “suffering” here), but not internal positive value (e.g., “pleasure” or “happiness” or “joy” or “Fun” or whatever label we want to use)?
I suppose it was because the original quote started with a negative framing, the assumption that the baby might not be fed.
I think both birth and death are stressful experiences that are not worth going through unless there are compensating other factors. I don’t think infants have enough of those if they die before they grow up.
Also I suspect human life is generally overrated, and the positives of life are often used as an excuse to justify the suffering of others. I do not trust people to make a realistic estimate and act with genuine benevolence.