[David Cooper]

I'll show another variation of trolley problem, where the one sacrificed for the five was a relative or romantic partner. Survey data shows that respondents are much less likely to be willing to sacrifice their life.

People who are emotionally attached to some of the individuals involved in a situation of that kind cannot be expected to make impartial moral judgements, or if they can, they can't be expected to apply them impartially - it is not wrong for someone to save someone they care about at the expense of more valuable people, but it is heroic if they don't. It's possible in many such situations that they'd be so tortured by making the "right" decision due to the loss that their own trauma would be bigger than that of all the relatives and friends of the other victims if the "wrong" decision was made, which could make the "wrong" decision potentially right. Fortunately though, the intelligent machines that we want to be making moral decisions will not have such biases and will not be traumatised at making rational decisions, so they will make be able to make correct decisions in all cases.

[David Cooper]

Let's assume that there are no uncertainty about all of those assumptions. At a glance, it seems to be obvious that the doctor should kill that tourist and provide his healthy organs to those five dying persons and save their lives.

No it doesn't - it is immediately obvious that one of the ill people can be sacrificed instead. However, you can introduce more information to rule that out - the healthy traveller's organs are compatible with all the others, but none of the others are compatible with each other. We now have a restored dilemma in which killing one person saves more. (This ignores organ rejection and decline - most transplanted hearts will fail within a decade, for example, but let's imagine that there's no such problem.

One of the important factors here is that no one wants to live in a world where they could be killed in such a way to save the lives of ill people (who wouldn't want to be saved in such a way either) - it's bad enough that you could die in accidents caused by factors outside of anyone's control, but you don't want to live in fear that you'll be selected for death to mend other people who may be to blame for their own medical problem or who may have bad genes which really shouldn't be passed on. You also don't want the fact that you've been careful to stay as healthy as possible to turn you into a preferred donor either - that could drive people to live unhealthy lives as it might be safer to risk being someone who needs a transplant than to be a good organ donor. However, if people's own morality is taken into account, it would serve someone right if they were used in this way if they've spent their life abusing others. As with all other moral issues, you have to identify as many factors as possible and then weight them appropriately so that the best outcome is more likely to be produced. A lot of the data needed to make ideal decisions isn't available yet though - it would take a lot of studying to find out how people feel in such situations and afterwards so that the total amount of harm can be counted up.

[David Cooper]

Since universal moral standard concerns about long term results, it would take a lot of factor to calculate, which might not make it practical. Bad results might come before the decision is made due to long duration of the calculation, and the factors influencing the calculation might have change before the calculation is complete.
Hence we need to create shortcut, rule of thumb, or hash table to deal with frequently occurring situations. They must be reasonably easy to calculate and work in most cases.

That's correct. When machines calculate the best course of action, they will apply all the most important factors in order of importance, getting better answers after each factor has been factored in, so when it reaches the point where they have to act, they'll do the best they can in the time available. It's harder for people to process the data in that way, but they will still have to make rapid decisions in many cases and will likely make a lot of bad decisions. If a set of simple rules can improve their performance, those rules should be used even if they're far from ideal. Different individuals should maybe have their own set of rules designed for them by intelligent machines, modifying them over time as they gain in experience, and as their mental powers decline too. AGI (the artificial general intelligence in future machines) will be able to design all those rules for individual people, so it isn't something we need to do directly - the important task for us is to make sure AGI is able to calculate morality correctly.

[hamdani yusuf (OP)]

Let's assume that there are no uncertainty about all of those assumptions. At a glance, it seems to be obvious that the doctor should kill that tourist and provide his healthy organs to those five dying persons and save their lives.

No it doesn't - it is immediately obvious that one of the ill people can be sacrificed instead. However, you can introduce more information to rule that out - the healthy traveller's organs are compatible with all the others, but none of the others are compatible with each other. We now have a restored dilemma in which killing one person saves more. (This ignores organ rejection and decline - most transplanted hearts will fail within a decade, for example, but let's imagine that there's no such problem.

One of the important factors here is that no one wants to live in a world where they could be killed in such a way to save the lives of ill people (who wouldn't want to be saved in such a way either) - it's bad enough that you could die in accidents caused by factors outside of anyone's control, but you don't want to live in fear that you'll be selected for death to mend other people who may be to blame for their own medical problem or who may have bad genes which really shouldn't be passed on. You also don't want the fact that you've been careful to stay as healthy as possible to turn you into a preferred donor either - that could drive people to live unhealthy lives as it might be safer to risk being someone who needs a transplant than to be a good organ donor. However, if people's own morality is taken into account, it would serve someone right if they were used in this way if they've spent their life abusing others. As with all other moral issues, you have to identify as many factors as possible and then weight them appropriately so that the best outcome is more likely to be produced. A lot of the data needed to make ideal decisions isn't available yet though - it would take a lot of studying to find out how people feel in such situations and afterwards so that the total amount of harm can be counted up.

Thanks for contributing to this discussion. I agree with most of your post above, so I'll try to identify where we split opinions. It's likely that we took different assumptions.
That's why I opened the statements you quoted by removing uncertainty of the assumptions, and started the second sentence with "at a glance". Of course the decision might change when the assumptions are changed.
Apparently you reject the equality of life expectancy.
I admit that when proposing to sacrifice the ill people, I omitted the possibility that none of them is compatible with each other. That's because I considered that organ rejections are caused by immune system due to mismatched gene. Hence I made a hidden assumption that since the traveler's organs are compatible with all of those ill people, their organs are compatible with each other. The proposed plan is preferred because no one ends up worse than no action.
I only wanted to emphasize that we should be open minded and try to think out of the box to get the best results with minimum negative impact. Perhaps with future technology of 3D printed organs we would no longer have to deal with this kind of dilemma any more.

[hamdani yusuf (OP)]

Since universal moral standard concerns about long term results, it would take a lot of factor to calculate, which might not make it practical. Bad results might come before the decision is made due to long duration of the calculation, and the factors influencing the calculation might have change before the calculation is complete. Hence we need to create shortcut, rule of thumb, or hash table to deal with frequently occurring situations. They must be reasonably easy to calculate and work in most cases. Their applications should align with the spirit of universal moral standard. This comparison might be made retrospectively when the decision has already been made before the calculation based on universal moral standard is finished. When they are in conflict, some exception should be made to the application of those shortcut rules.

Biological evolution has provide us with a basic and simple shortcut rule, which is to avoid pain. This can be done through reflex which is very fast since it doesn't involve central nervous system. A little bit more complex rules are our instinct to seek for pleasure and to avoid suffering. I  think hedonism and utilitarian are confusing the tool with the goal.

[alancalverd]

The mathematical resolution of the simplest trlley problem assumes that your universal moral standard is to maximise the number of live humans. Since this will inevitably lead to the starvation of our descendants, it is a questionable basis for ethics.

[David Cooper]

Biological evolution has provide us with a basic and simple shortcut rule, which is to avoid pain. This can be done through reflex which is very fast since it doesn't involve central nervous system. A little bit more complex rules are our instinct to seek for pleasure and to avoid suffering. I  think hedonism and utilitarian are confusing the tool with the goal.

I can't follow that. What's the tool there and what's the goal?

[David Cooper]

The mathematical resolution of the simplest trlley problem assumes that your universal moral standard is to maximise the number of live humans. Since this will inevitably lead to the starvation of our descendants, it is a questionable basis for ethics.

The assumption is that saving more people is better. That's quite different from saying that there should be more and more people until everyone starves. For your objection to be relevant, you'd have to set the trolley problem in a place where people are starving and a cull would be a good idea, but you'd want that cull to take out the most immoral people rather than people who have likely been tied to a track by the most immoral person. So, you minimise the number of people killed by the trolley, then you count how many people still need to be culled and you take out that number of the most immoral ones.

The only place I've heard the idea that we should maximise the number of live humans even if that leads to them all being right on the edge of starvation is with the Mere Addition Paradox, but that paradox contains a major mathematical error which renders it a pile of nonsense.

[hamdani yusuf (OP)]

The mathematical resolution of the simplest trlley problem assumes that your universal moral standard is to maximise the number of live humans. Since this will inevitably lead to the starvation of our descendants, it is a questionable basis for ethics.

A lot of disputes may arise if we don't agree with the scope of the subject of discussion. I've stated that universal moral standard is not limited as narrow as the existence of human beings, as long as there are conscious beings. It should have been in place before modern human exist, and it should still be in place when human has evolved into other species. as long as there exist conscious beings.
Universal moral standard concerned with the result in the long run, which means it covers an extended time scale.
If we have 10 billion people living happily in 1 generation but then go extinct in the next generation, it doesn't fulfill the goal of universal moral standard.

[hamdani yusuf (OP)]

Biological evolution has provide us with a basic and simple shortcut rule, which is to avoid pain. This can be done through reflex which is very fast since it doesn't involve central nervous system. A little bit more complex rules are our instinct to seek for pleasure and to avoid suffering. I  think hedonism and utilitarian are confusing the tool with the goal.

I can't follow that. What's the tool there and what's the goal?

The goal is what is preferred in the long run. The rules used as the shortcut is the tool.

[alancalverd]

No matter what the species or timescale, if maximisation of the number of living organisms is the prime objective and it has the unfettered capacity to maximise, it will eventually run out of food or poison itself with its own excrement. Never mind humans, you can observe the endpoint with lemmings and yeast (which is why wine never exceeds 20% alcohol).

The consequence of "saving more people" as an ethical axiom is visible in a developed society where the National Health Service and HM Prison Service are forced by the courts to expend resources extending the lives of people who want to die, and in less developed countries where public health and tradition have increased the population to the point that it cannot be sustained by an inherently marginal agriculture.

"Do as you would be done by" looks like a more generally applicable motto, but the fact that it can't be applied to the trolley problem suggests that there may not be a single universal moral standard. And here's where my thinking became suddenly heretical and digressive:

In the absence of a universal principle, we often choose an arbitrary standard. "The man on the Clapham omnibus" serves for many legal questions but some people revert to a single figure and ask "what would Jesus do?" Sitting here, my first thought was "well, he wouldn't eat pork" (I've been refereeing a medical experiment that involves eating a standard fatty meal)...and then (apropos lemmings, I suppose) I wondered about the Gadarene swine. Who was herding pigs in Israel?

Anyway, returning to the relative value question, I had a fine example of this in my days in Civil Defence. An enthusiastic young lecturer from the Home Office was explaining post-nuclear-strike policy to our village command. The obvious priority, he said, was to feed and protect the elderly, pregnant women, and children. My boss, who had  actually led his artillery regiment in conflict, said that with at least 60% of the population dead or dying and an imminent threat of invasion, his priority was to feed and protect men aged 16 - 60 who could dig graves and fight. "With any luck, we will survive to make children and pensioners later, but the converse is impossible".

[hamdani yusuf (OP)]

No matter what the species or timescale, if maximisation of the number of living organisms is the prime objective and it has the unfettered capacity to maximise, it will eventually run out of food or poison itself with its own excrement. Never mind humans, you can observe the endpoint with lemmings and yeast (which is why wine never exceeds 20% alcohol).

I think you might want to revisit my answer to the what, who, when, where, why, and how questions about morality in post #9, #18, #29, #30, #33,#35, #39, #40, #41, #45 - $48.

[David Cooper]

Biological evolution has provide us with a basic and simple shortcut rule, which is to avoid pain. This can be done through reflex which is very fast since it doesn't involve central nervous system. A little bit more complex rules are our instinct to seek for pleasure and to avoid suffering. I  think hedonism and utilitarian are confusing the tool with the goal.

I can't follow that. What's the tool there and what's the goal?

The goal is what is preferred in the long run. The rules used as the shortcut is the tool.

So when you say utilitarianism is is confusing the tool with the goal, how is it confusing a shortcut with what's preferred in the long run? Where's the incompatibility between the two?

[David Cooper]

No matter what the species or timescale, if maximisation of the number of living organisms is the prime objective and it has the unfettered capacity to maximise, it will eventually run out of food or poison itself with its own excrement. Never mind humans, you can observe the endpoint with lemmings and yeast (which is why wine never exceeds 20% alcohol).

If you've got things set up properly, everything works in cycles: composting toilets convert the worst kind of waste into beautiful soil for growing new food in - it all works fine so long as the sun's putting energy in. That kind of requirement has to govern the maximising of the population so that you don't get to a point where everyone drowns in sewage. It's about reaching a maximum stable population so that you don't end up like animals where they go through repeated population crashes as the food supply fluctuates. Of course, maximising a stable population isn't idea either if that reduces quality of life due to lack of other resources which can be essential for well-being, such as living in a pleasant environment rather than all being crammed together in filthy towns.

"Do as you would be done by" looks like a more generally applicable motto, but the fact that it can't be applied to the trolley problem suggests that there may not be a single universal moral standard. And here's where my thinking became suddenly heretical and digressive:

Who says it can't be applied to the trolley problem? Of course it can. If you're one of the people tied to the track, you want the person to make the trolley go the other way, and if you're the one by the lever, you want to make the trolley go the other way for the person tied to the track. You can either let down one person or five, so your moral duty is to let down one rather than five.

In the absence of a universal principle, we often choose an arbitrary standard. "The man on the Clapham omnibus" serves for many legal questions but some people revert to a single figure and ask "what would Jesus do?" Sitting here, my first thought was "well, he wouldn't eat pork" (I've been refereeing a medical experiment that involves eating a standard fatty meal)...and then (apropos lemmings, I suppose) I wondered about the Gadarene swine. Who was herding pigs in Israel?

The problem we have is not in finding the universal principle, but in getting stupid people to recognise that they should be accepting it. Many people only respect authority as they're incapable of thinking for themselves, but we have lots of fake authorities. That messes everything up, and it causes all manner of conflicts and other abuses. Religions and faulty ideologies are the main barrier to progress.

[hamdani yusuf (OP)]

Biological evolution has provide us with a basic and simple shortcut rule, which is to avoid pain. This can be done through reflex which is very fast since it doesn't involve central nervous system. A little bit more complex rules are our instinct to seek for pleasure and to avoid suffering. I  think hedonism and utilitarian are confusing the tool with the goal.

I can't follow that. What's the tool there and what's the goal?

The goal is what is preferred in the long run. The rules used as the shortcut is the tool.

So when you say utilitarianism is is confusing the tool with the goal, how is it confusing a shortcut with what's preferred in the long run? Where's the incompatibility between the two?

To answer your question, I need first to continue my assertion about progress of increasing complexity of shortcut rules provided by biological evolution. With increasing complexity, more factors can be included in the calculation to generate actionable output. More complex rules can accommodate more steps into the future and at some point, they appear as planned actions.
We can put some milestones in the continuum of complexity of shortcut rules. The next step from instinct is emotion. Emotion includes anticipation of near future events. We can feel sad/happy/fear/angry before events which potentially cause pleasure/pain actually happens.

[jimbobghost]

I have attempted to follow the wisdom of the posters herein.

I hoped to find an answer to the question.

unfortunately, I must answer the problem with a simple conclusion:

no, there is no "universal moral standard".
there can never be a "universal moral standard" until every sentient species in the universe agrees upon the standard of the combined species.

sorry.

[David Cooper]

unfortunately, I must answer the problem with a simple conclusion:

no, there is no "universal moral standard".
there can never be a "universal moral standard" until every sentient species in the universe agrees upon the standard of the combined species.

They don't need to agree on it - correct morality can be imposed on them regardless, and will be. If you have to live the life of a human, a dog, an alien and a cat, and if all four of these animals cross paths in some way, their interactions can be guided by a simple method. How should you as the human treat the others? How should you as the alien treat the others? As the dog and the cat, you won't have the wit to work out how to treat the others morally, but as the human and alien, you can. An AGI system can work out what's best for the dog and the cat to do too, so it can intervene to protect them from each other, just as the human and alien can intervene. Because you are to imagine living all four lives, you don't want the alien to enslave the human because you will gain less from that as an alien than you lose as the human. You don't want the dog to kill and eat the cat either for the same reason, and it doesn't matter that they can't understand morality.

[jimbobghost]

David,

as always, your wisdom is awesome.
pardon me if I inject a bit of humor here:

what do you have against eating pussies?

[David Cooper]

what do you have against eating pussies?

Roof rabbits (cats) certainly have their place in times of war when food's short, but if someone fancies a bit of ***** (edit: crikey - it doesn't like the singular of pussies!) at other times, they really need to identify the owner and enter careful negotiations with them.

(Test: beaver...)

[jimbobghost]

amazing...first time I have noted the resident "hall monitor" restrict a word/phrase.

I have been impressed with the latitude allowed herein.