Mild welfarism: avoiding the demandingness of total(itarian) welfarism

For the full, academic paper, see here.

The demandingness of total welfarism

Welfare measures how good a situation is for an individual. So we value welfare, and everyone’s welfare counts equally. According to the moral theory of total welfarism, we should choose the option (situation) that maximizes the total welfare, i.e. the sum of the welfare values of everyone who exists, existed or will exist. Total welfare is the objective that we have to maximize. The more welfare the world contains, the better that world is.

Probably the biggest objection to total welfarism, is its demandingness. The theory entails huge sacrifices: we have to give away our money and resources to the poorest people until we become as miserable as the poorest person and live lives at the subsistence level, we have to donate our kidneys and other organs for transplantation until an organ shortage in the hospitals is eliminated, we have to donate blood until we become anemic, we have to have more children than we would like until an extra child gets a zero or negative welfare, we have to prioritize helping the poorest children instead of our own children, we have to sacrifice everything in order to cause the existence of huge numbers of extra people in the far future even when those extra people will have lives barely worth living, we have to help all present and future animals and other sentient beings as much as we would help our closest friends and family members, we have to perform medical experiments on our bodies to find cures for other people, we have to save and invest almost all of our income for the benefit of future generations, we have to spend most of our time figuring out how to improve total welfare, and so forth. It is clear that no-one lives according to the demands of total welfarism. This moral theory, which we can rather call totalitarian welfarism, is too counter-intuitive. Can we change this theory to avoid its demandingness?

A modified theory: mild welfarism

Why is total welfarism so demanding? Because we have to take into consideration the welfare of everyone else. If the welfare of other people matters, we may have to sacrifice ourselves to increase the welfare of others. If only we could discount the welfare of others, we avoid the sacrificial demandingness problems of total welfarism. So what if we could say that someone else’s welfare gain doesn’t count? Of course, if I may say that your welfare doesn’t matter, then you may do the same for my welfare. Hence, what if everyone has a right to discount the welfare of other people? Discounting the welfare of a person in a situation means deducting an amount from the welfare value that person has in that situation. Instead of taking the sum of the welfare values of all individuals, we take the sum of the welfare values minus the discounts. The best option is the one with the highest total discounted welfare, i.e. the highest sum of welfare minus discounts.

If you don’t want to make a sacrifice for a welfare gain of someone else, you can simply discount that welfare gain. Of course, in many cases, a person whose welfare is discounted by you, can reasonably object to that discounting. If your welfare is discounted, you are likely getting a lower welfare compared to the total welfarist choice for the situation that maximizes the sum of undiscounted welfare. As you do not want your welfare being discounted, we should not simply give everyone the right to discount the welfare of others. A restriction of this right is needed.

As a restriction, we can say that everyone has the right to discount the welfare of other individuals as long as those discounted individuals cannot validly object or complain against their welfare being discounted. But what counts as a valid objection or complaint? When is a complaint invalid? The right to discount welfare always involves two parties: the discounting person who discounts another person’s welfare and the discounted person whose welfare is being discounted. The fact that a right involves a relationship between two individuals means that there are two ways in which a complaint becomes invalid. To check the validity of a complaint, we have to ask the question: what if either of these two individuals (the discounting and the discounted person) did not exist? Would the person whose welfare is discounted be better-off?

Suppose there are two people, An, the agent who makes a choice to benefit Ben, and Ben, the beneficiary who may be benefitted by An. And there are two possible options for An, called Sacrifice and Non-Sacrifice. In Sacrifice, An sacrifices part of her welfare for the benefit of Ben. If Sacrifice generates the highest total welfare, this option should be chosen according to total welfarism. To avoid this selection of Sacrifice, An discounts the welfare of Ben in the option Sacrifice. The welfare of An plus the discounted welfare of Ben is lower in option Sacrifice than the welfare of An plus Ben in option Non-Sacrifice.

So An chooses Non-Sacrifice, as this maximizes the total discounted welfare. In Non-Sacrifice, Ben is not benefitted. Can he complain, and is his complaint valid? Ben’s complaint is valid only if first, he exists, and second, his welfare gain could also be achieved if An did not exist. Hence, two conditions need to be satisfied: the existence of Ben is necessary and the existence of An is not necessary for Ben’s welfare gain.

If everyone has the right to discount welfare under the restriction that no-one can validly complain against the discounting, then instead of choosing the option that maximizes the total welfare, we can choose the option that maximizes the total discounted welfare. Choosing the latter option is less demanding than choosing the option that maximizes the total welfare. Hence, this modified version of total welfarism can be called ‘mild welfarism’, as it is less demanding than total welfarism.

The next two sections discuss why mild welfarism is less demanding. The two sections refer to the two conditions of validity of complaints.

The first condition becomes important in cases where our choices determine the existence of future people. If we can choose whether or not someone will exist in the future, that person is a possible person, i.e. someone who does not exist in all possible futures that we can choose. In contrast, if a person exists in all possible options that one can choose, that person is a necessary person.

Here we enter population ethics, the branch of ethics that studies which situation is the best or which option we should choose when different situations or options contain different existing people. Which future state of the world should we choose when our choice determines who will exist in the future? How should we treat the welfare of possible or future people who may not come into existence?

The second condition relates to a right not to be abused, i.e. a right not to be used as a means against one’s will for the benefit of someone else.

Population ethics

The first condition seems trivial: if Ben does not exist, of course he cannot complain. But this condition has interesting consequences. The case becomes interesting when Ben does not exist in either Sacrifice or Non-Sacrifice. If Ben exists in only one of those two possible options, then the choice of An to sacrifice herself determines the existence of Ben. An can choose to bring Ben into existence, which means that before An decides to bring Ben into existence, Ben is a possible person: there are futures in which Ben does not exist.

Suppose Ben does not exist in the option Non-Sacrifice, but he exists and lives a life with a positive welfare in the option Sacrifice. An has a lower welfare in option Sacrifice than in option Non-Sacrifice. Therefore, An does not want to sacrifice herself to bring Ben into existence. So An can decide to discount the welfare of Ben in Sacrifice. An can deduct an arbitrary amount from the welfare value that Ben gets in option Sacrifice, such that Ben’s discounted welfare becomes zero or negative. If Ben cannot complain against his welfare being discounted in option Sacrifice, the discounting is valid. According to the theory of mild welfarism, An is allowed to choose option Non-Sacrifice because the welfare of An in option Non-Sacrifice is larger than the sum of the welfare of An and the discounted welfare of Ben in option Sacrifice.

So the question becomes: can Ben complain against his welfare being discounted in option Sacrifice? No, because if option Sacrifice were chosen, Ben achieves the highest welfare, which is positive. There is no other available option where Ben would be better-off. His life is worth living, so he has no reason to complain against the selection of Sacrifice. And if the option Non-Sacrifice were chosen, Ben cannot complain either because he does not exist. Hence, the positive welfare of possible or future people can be fully discounted. If Ben has a positive welfare in option Sacrifice and does not exist in option Non-Sacrifice, then there exists no-one in option Non-Sacrifice who can say that Non-Sacrifice is a worse option than Sacrifice. In that sense, we can say that Non-Sacrifice is not worse than Sacrifice. If it is not worse, then An is allowed to choose option Non-Sacrifice.

In general, people have the right to discount the welfare of others in some options if the option with the highest total discounted welfare is one in which the people with the discounted welfare do not exist. If the discounted people do not exist in the chosen option with the highest total discounted welfare, the existence of the discounted people is in a sense not necessary, and in that case the complaint against discounting becomes invalid.

With the right to discount the positive welfare of possible people, we can avoid some demandingness issues of total welfarism. For example, we do not have to give birth to more children than we would like and we do not have to sacrifice everything in order to cause the existence of huge numbers of extra people in the far future even when those extra people have lives barely worth living, with a positive but very small welfare.

This theory of mild welfarism has important implications for our consideration of future generations. Consider an existential catastrophe that kills everyone. Such a disaster is bad for the existing people who experience the catastrophe, but it is not bad for the quadrillions of non-existing future people who could have had extremely satisfying lives if the catastrophe did not occur. If we could avoid the catastrophe but choose not to, the non-existent future generations cannot complain against us letting the catastrophe occur. In the theory of mild welfarism, such an existential catastrophe is bad, but it is less bad than in the theory of total welfarism, because the latter theory includes the huge welfare loss of the non-existing future happy people.

The theory of mild welfarism has also something interesting to say in the example of happy animal farming. Is it permissible to bring into existence farm animals that are overall happy, but are prematurely killed for their meat? The human consumer who enjoys eating meat gains welfare, and the farm animal has a positive (but perhaps small) welfare. But once the farm animal is brought into existence, it also becomes possible not to kill that animal, but to take care of that animal for example at an animal sanctuary. So we have three options: No-farming (in which the animal is not brought into existence), Farming (in which the animal is brought into existence, lives a slightly happy life on a farm but is prematurely killed and eaten by a human) and Sanctuary (in which the animal is brought into existence and lives a very happy and long life because she is taken care of by a human).

Choosing option Sanctuary, the human can no longer enjoy eating the meat of that animal, but instead has to sacrifice time and resources to help the animal at the cost of her own welfare. In the animal sanctuary option, the human may have a lower welfare than in the situation No-farming where the animal was never brought into existence.

The human prefers option Farming. If that situation has the highest sum of welfare, that situation can be chosen. But we can expect that the situation Sanctuary has the highest sum of welfare, as the animal is much better-off and the human is only slightly worse-off than in option Farming. According to total welfarism, the human should choose option Sanctuary. In extremis, total welfarism could come to the conclusion that the human should choose an option Large Sanctuary, where the human breeds many happy animals and gives up a lot of time and resources to take care of those animals on a large animal sanctuary. The large welfare of the many animals trumps the huge welfare loss for the human. The human has to make a huge sacrifice for the animals.

For the human, option Sanctuary (and definitely Large Sanctuary) is the least preferred, as the human has the lowest welfare in that option. To avoid option Sanctuary, the human can discount the welfare of the animal in option Sanctuary, such that the sum of the welfare minus discount in option Sanctuary is lower than the sum of the welfare of the human and the animal in option Farming. However, this does not yet justify the selection of option Farming, because in that option, the animal exists. And as the animal in option Farming has a lower welfare than in option Sanctuary, the animal in option Farming can complain against the discounting of its welfare in option Sanctuary. Hence, discounting the animal’s welfare in option Sanctuary is not valid. But the human can still avoid the selection of option Sanctuary, by also discounting the animal’s welfare in option Farming. If the welfare of the animal in both options Farming and Sanctuary is sufficiently discounted, option No-farming becomes the option with the highest total discounted welfare (the highest sum of welfare minus discounts). The selection of option Sanctuary is avoided at the cost of option Farming becoming impermissible. The only permissible option is No-farming.

Neither the human nor the non-existing farm animals can complain against this choice of No-farming. Hence, with this mild welfarism theory, animal farming is not permissible, not even when the farm animals would have a positive welfare (unless they would not complain against their welfare in option Sanctuary being discounted, for example they would have such a high welfare in option Farming that they would never complain against being used for their meat). And of course unhappy animal farming is definitely not permissible, because the farm animal has a negative welfare and is used against her will as a means to someone else’s ends. The latter brings us to cases where the second condition is violated: cases where the existence of a person (e.g. the animal) is not necessary for the welfare gain of someone else (e.g. the meat eater).

The right not to be used as a means against one’s will

There is a correspondence between the right to discount the welfare of others and the right not to be used as a means against one’s will. We can see this by considering some examples.

As a first example, suppose An and Ben are on an island and they find some food. An takes all the food. If An does not give some food to Ben, Ben will die from starvation. An can make a sacrifice by giving some food to Ben. But An wants to keep all the food for herself: no sacrifices. To justify her choice for the option Non-Sacrifice, she fully discounts the welfare gain that Ben can get by giving him some food. Ben’s welfare is discounted, and he complains. That complaint is valid, because Ben exists (first condition), and the existence of An is not necessary for Ben having access to food (second condition). On the contrary, if An did not exist, Ben could have all the food. As Ben’s complaint is valid, An is not allowed to discount Ben’s welfare and hence she should not take all the food for herself.

Another example: suppose Ben has a kidney disease and needs a kidney of An in order for him to survive. An does not want to sacrifice her kidney and discounts the welfare gain that Ben can get from an organ transplantation. Can Ben complain against this discounting? The first condition for a valid complaint is satisfied: Ben exists. But the second condition is not satisfied. Ben would die if An did not exist, because if she did not exist, there is no kidney available for transplantation to save Ben. The presence of An is necessary for the welfare gain (the survival) of Ben. It is this welfare gain that can be validly discounted by An. If An would discount this welfare gain of Ben, Ben cannot validly complain against that discounting, because Ben could not have gained that welfare in the absence of An. If Ben says to An: “You should give me your kidney”, An can respond by saying: “No, I do not have a duty to give you my kidney, because if I did not exist, you would not have my kidney either, so you would not have been better-off.” Note that such a response is not valid in the previous example of the island. If Ben says to An: “You should give me some food”, An cannot respond by saying: “No, I do not have a duty to give you food, because if I did not exist, you would not have the food either.” The fact that the existence of An is necessary for the welfare gain of Ben in the organ transplantation example, makes this example different from the island example.

To save Ben in the organ transplantation case, An has to exist. Or in other words: to choose option Sacrifice, An has to exist. If An did not exist, she cannot sacrifice herself. If in the absence (or non-existence) of An it is not possible to make Ben better-off than in option Non-Sacrifice, it means that the existence of An is necessary to improve the welfare of Ben. The welfare that Ben gains when switching from Non-Sacrifice to Sacrifice can only be achieved if An exists. When Ben could not be made better-off in the absence of An, the right of An to discount the welfare gain of Ben is valid, because Ben cannot validly complain against his welfare gain being discounted.

In general, people have the right to discount the welfare of others in some options if those options are not possible if all the persons who do the discounting did not exist. An option is not possible if one cannot act in a way that generates the same outcomes (for everyone except the group of people who were assumed not to exist) of that option. If an option is only possible if the discounting people are present, the existence of those people is in a sense necessary for that option, and in that case the complaint against discounting becomes invalid.

If the presence or existence of An is necessary for the benefit of Ben, we can say that An is a means for the benefit of Ben. If An benefits Ben, then Ben uses An as a means. A hammer is a means to hit a nail, because if the hammer did not exist, you could not hit the nail with it. If the use of An as a means involves an unwanted sacrifice of An, i.e. she does not want to get a lower welfare, then she does not want to be used as a means. In that case, using her is against her will. It is abusing her. To protect An against abuse, we can say that An has the right not to be used as a means against her will for the benefit of someone else. Or, in the example of organ transplantation, An has the right to bodily autonomy, that her body is not used as a means against her will. This right is actually a consequence of her more fundamental right to discount the welfare gains of other people if those people cannot validly complain.

More ideas follow from this fundamental right. For example the idea that ‘Do no harm’ is more important than ‘Offer help’: your existence is not necessary for you not to cause harm, whereas you can only offer help if you exist. Resulting from this, doing harm is worse than allowing harm or not offering help. And when we offer help, we are allowed to be partial and prioritize those who are most dear to us, because if we have to help strangers when we prefer to help our friends or family instead, those strangers use us as merely a means.

As an example of the latter issue, consider a burning house dilemma where you can save either your child locked up in one room or two other, unknown children locked up in another room. You only have time to open one door. Total welfarism says that you should save the two unknown children. You do not want that, so you discount the welfare gains that those two children would get if you save them. The children cannot validly complain against their welfare gains being discounted, because if you did not exist, those children could not have been saved and would not have been better-off than when you decide to save your own child. So it is permissible for you to save your own child. It is permissible to be partial towards people you hold dear. You are allowed to spend your money on a birthday gift for your child instead of donating that money to a charity that saves children in a poor country.

An upper-bound on the amount of discounting

Without the right to discount welfare, we have total welfarism that is too extreme in the sense of being too demanding. However, if the right to discount the welfare of people is absolute or infinite in strength, we have no duties at all to help others and we can completely neglect the positive welfare of possible or future people. This may be too extreme in the other direction. The pendulum swings too far in the opposite direction of non-demandingness.

In an intermediate position, everyone has a bounded right to discount the welfare gains of others if the discounting cannot be validly objected. The right is finite in strength. There is an upper-bound on the amount of permissible discounting. That means we do have some duty to help others if it is at a sufficiently large benefit for the beneficiaries and a sufficiently small cost of our own welfare. We do have to be altruistic (and impartial) to some degree. We do have non-zero duties to guarantee the existence of (larger populations of) happy future generations. In general, when the total welfare in one option is sufficiently much higher than in the other options, one should choose that option. That means the total amount of discount should be smaller than the gain in total welfare one gets by choosing the option with the highest total welfare.

This theory of mild welfarism has a free parameter, the maximum amount of permissible discounting, that measures the maximum strength of the right to discount welfare. When the parameter is zero, we end up with total welfarism (no discounting is permissible), and when it is infinite, we end up with unbounded discounted welfarism.

So how strong is this discounting right? We can make it simple, by maximizing the autonomy of moral agents. For individual choices, moral agents or decision-makers are free to choose for themselves how high they set the bar, as long as it is not infinite. They have to pick a finite upper-bound on the discount. They can even choose different upper-bounds in different contexts, dilemmas or situations. And when it comes to collective choices, a democratic consensus procedure could be applied to determine the upper-bound on permissible discounting.

Summary

I proposed a modification of total welfarism that completely avoids the demandingness problem in a manner that causes the least amount of complaints. According to mild welfarism, we should choose the option that maximizes the total validly discounted welfare, i.e. the sum of everyone’s welfare minus the complaint-free discounts. Everyone has a bounded or limited right to discount the welfare of others if the welfare discounted people cannot validly complain against their welfare being discounted. The discounting is valid only if the complaints are invalid. A complaint is valid only if that welfare discounted person exists and the welfare gain of that person could also be achieved if the people who discount that person’s welfare did not exist.

In general, everyone has the right to discount someone else’s welfare in some options if those options are not possible (or if the outcomes of those options would change) if the people who do the discounting did not exist or if the option with the highest total discounted welfare (the sum of everyone’s welfare minus discounts) is one in which the people whose welfare is discounted do not exist. A complaint made by welfare discounted people is not valid if the existence of the discounting people is in a sense necessary (the option cannot be chosen if none of the discounting people existed) or if the existence of the discounted people is in a sense not necessary (the option that should be chosen is one where the people do not exist).

For individual choices, people can freely choose for themselves a finite upper-bound on the amount of discounting. For collective choices, this upper-bound can be decided democratically.