People often criticise me of thinking of potential people as Steven Landsburg describes without necessarily endorsing:
…like prisoners being held in a sort of limbo, unable to break through into the world of the living. If they have rights, then surely we are required to help some of them escape.
Such people seem to believe this position is required for considering creating good lives an activity with positive value. It is not required, and I don’t think of potential people like that. My position is closer to this:
Benefit and harm are comparative notions. If something benefits you, it makes your life better than it would have been, and if something harms you it makes your life worse than it would have been. To determine whether some event benefits or harms you, we have to compare the goodness of your life as it is, given the event, with the goodness it would otherwise have had. The comparison is between your whole life as it is and your whole life as it would have been. We do not have to make the comparison time by time, comparing each particular time in one life with the same time in the other life.
That is John Broome explaining why death harms people even if they hold that all benefit and harm consists of pleasure and pain, which are things that can’t happen when you are dead. The same goes for potential people.
Yes, you can’t do much to a person who doesn’t exist. They don’t somehow suffer imaginary pains. If someone doesn’t exist in any possible worlds I agree they can’t be helped or harmed at all. What makes it possible to affect a potential person is that there are some worlds where they do exist. It is in the comparison between these worlds and the ones where they don’t exist where I say there is a benefit to them in having one over the other. The benefit of existing consists of the usual things that we hold to benefit a person when they exist; bananas, status, silly conversations, etc. The cost of not existing relative to existing consists of failing to have those benefits, which only exist in the world where the person exists. The cost does not consist of anything that happens in the world where the person doesn’t exist. They don’t have any hypothetical sorrow, boredom or emptiness at missing out. If they did have such things and they mattered somehow, that would be another entirely separate cost.
Often it sounds crazy that a non-existent person could ‘suffer’ a cost because you are thinking of pleasures and pains (or whatever you take to be good or bad) themselves, not of a comparison between these things in different worlds. Non-existent people seem quite capable of not having pleasures or pains, not having fulfilled preferences, not having worthwhile lives, of not having anything at all, of not even having a capacity to have. Existent people are quite capable of having pleasures (and pains) and all that other stuff. If you compare the two of them, is it really so implausible that one has more pleasure than the other?
‘Potential people’ makes people think of non-existing people, but for potential people to matter morally, it’s crucial that they do exist in some worlds (in the future) and not in others. It may be better to think of them as semi-existing people.
I take it that the next counterargument is something like ‘you can’t compare two quantities when one of them is not zero, but just isn’t there. What’s bigger, 3 or … ?’ But you decide what quantities you are comparing. You can choose a quantity that doesn’t have a value in one world if you want. Similarly I could claim all the situations you are happy to compare are not comparable. Getting one hundred dollars would not benefit you, because ‘you without a hundred dollars’ just won’t be around in the world where you get paid. On the other hand if you wanted to compare benefits to Amanda across worlds where she may or may not exist, you could compare ‘how much pleasure is had by Amanda’, and the answer would be zero in worlds where she doesn’t exist. Something makes you prefer an algorithm like ‘find Amanda and see how much pleasure she has got’, where you can just fail at the finding Amanda bit and get confused. The real question is why you would want this latter comparison. I can see why you might be agnostic, waiting for more evidence of which is the true comparison of importance or something, but I don’t recall hearing any argument for leaping to the non-comparable comparison.
Image via Wikipedia
In other cases it is intuitive to compare quantities that have values, even when relevant entities differ between worlds. Would you say I have no more orange juice in my cup if I have a cup full of orange juice than if I don’t have a cup or orange juice? I won’t, because I really just wanted the orange juice. And if you do, I won’t come around to have orange juice with you.
I have talked about this a bit before, but not explained in much detail. I’ll try again if someone tells me why they actually believe the comparison between a good life and not existing should come out neutral or with some non-answer such as ‘undefined’. Or at least points me to where whichever philosophers have best explained this.
Meteuphoric 
“In other cases it is intuitive to compare quantities that have values, even when relevant entities differ between worlds. Would you say I have no more orange juice in my cup if I have a cup full of orange juice than if I don’t have a cup or orange juice? I won’t, because I really just wanted the orange juice.”
This isn’t the analagous question.
Don’t ask how much more orange juice you have in one world than in the other. Instead, ask how much more orange juice is contained in the cup. That question has no meaningful answer, because in the second world there is no cup.
This is analgous to the question of whether there is more good in the world where the potential person exists versus the one where they don’t. If you value the person’s existence, then the first world has more orange juice, I mean utility. But if their existence is neutral to everyone but themselves, then there is no difference in utility between the two worlds – the first world does not have a greater amount of orange juice contained in the cup because there is no cup to contain it, and it has no more utility for the potential person because there is no person to have utility.
I should have said:
“The second world does not have a lesser amount of orange juice contained in the cup because there is no cup to contain it, and it has no less utility for the potential person because there is no person to have utility.”
My initial wording was a bit confused.
Eddie is assuming that the utility is measured in the present world. And that is wrong; when calculating expected utility, you compute the utility expected at some point in the future. When those people exist.
“‘Potential people’ makes people think of non-existing people, but for potential people to matter morally, it’s crucial that they do exist in some worlds (in the future) and not in others. It may be better to think of them as semi-existing people.”
On the contrary. Thinking of them as semi-existing people is factually incorrect, and obscures the most relevant fact about them, namely that they do not exist. That they might exist in the future does not change the fact that they do not exist now, and thus do not have any preferences.
If, due to choices we make today, it turns out that someday in the future they do come into existence, then at that point they will have preferences. But if we make a different choice and they never exist, they will never have had preferences and thus will never have suffered anything by virtue of not having had their preferences fulfilled.
If the fact that someone doesn’t exist now is relevant, do you also not care about people who don’t exist now, but will do? If so, I think you are unusual. Most people seem to claim that it’s possible to benefit future people. And as far as I know pretty much everyone agrees that you can’t benefit people who can never exist. So it seems the relevant feature of the people under debate is that they are in neither of these categories – their existence varies across worlds.
Be careful to distinguish three groups:
. Those who will never exist
. Those who could exist
. Those who will exist
I think we agree that those who will never exist deserve no consideration.
I think we also agree that those who will exist deserve as much consideration as any other person. I didn’t say this above, but I’m saying it here in response to your question.
But the key point regarding your entire post is how we feel about those in the middle group: the people who don’t exist yet, but could. Note especially that I don’t mean those who, say, have a 90% chance that they will exist and only a 10% change that they won’t, i.e. those for whom we are uncertain about their future existence. When I say “could” exist, I’m specifically referring to people whose existence we have the power to cause, and who we are deciding whether or not to create. Those are the people at issue in this discussion – the ones whose existence we are trying to determine the value of, so that we can argue for or against the morality of trying to bring that existence about.
You are arguing (I believe) that we should choose today to bring about future worlds in which potential people exist, because those worlds will have more value to those people than the worlds in which they don’t exist. I argue that those people cannot place a value on the worlds in which they do not exist – not even a value of zero – because in such worlds they do not have preferences.
Let me distinguish this from two other cases:
First, people who will exist: the comparisons are between worlds in which they exist but are relatively unhappy and those in which they will exist and are relatively happy. Yes, they do not exist now, and thus do not yet have preferences, but we should strive to satisfy the preferences they will have when they do exist. Unlike the merely potential people, they will have preferences in every future world.
Second, those who exist now but won’t exist in the future, i.e. those who will die. They may have preferences today about which future worlds should exist, even though in those future worlds they will no longer have any preferences. We should strive to bring about those future worlds – not to satisfy their future preferences, but to satisfy their present preferences. Potential people, like the soon-to-die, will not exist in (some) futures and thus will not have preferences in those futures. But unlike the soon-to-die, potential people have no preferences in the present either.
It’s a lot easier to defend your position through veil-of-ignorance type arguments. Consider two possible worlds, one in which person A tosses a coin to decide whether or not to have a child (called B), and one in which they do it definitely. Suppose further that B will have an exceptionally nice life, and that behind the veil of ignorance B is one of the people I think I might be. My chance of being B behind the veil-of-ignorance is in some sense higher in the second world than in the first, and since by assumption A’s child’s life is exceptionally nice, increasing the probability of being B increases my expected utility behind the veil.
If you don’t like the discussion of the “chance of being B”, think rather about the veil-of-ignorance as obscuring the properties of the life you will have. In the possible world in which A tosses a coin, the probability of having an “exceptionally nice” life must be lower.
The problem with veil of ignorance arguments is that it’s very hard to agree on a neutral reference set of possible people, i.e. on “prior beliefs” behind the veil. If the reference set is large enough to include all future beings, and my prior is on continued economic growth then introducing another person now (even one with an “exceptionally nice life”) may well reduce expected utility as it reduces the chance of being born in the far future.
I like your moral reasoning framework, although I think its incredibly unintuitive.
It seems to me you’re trying to move beyond talking about utility as it applies to the clasically and fundamental concept of a person, an entity which – as most people would conceptualise it – has a more or less solid boundary in space and time of a single, deterministic universe, and has some sort of mental qualities that we think are of moral concern. Rather, you’re concerned with utility of entities that exist over the phase space of all possible universes (there are various ways to formalise it, but I think the intuition is clear). This probabilistic approach is I think in general the only kind of robust reasoning that can be applied in the face of the huge uncertainty we have in our knowledge of the future (without even considering that physical reality itself might be constituted this way, ala Many Worlds Quantum.)
But our brains aren’t wired to reason especially well over even a universe with a single deterministic set of events in four dimensional spacetime (look at the muddle people get into over questions like free will etc), let alone the powerset of one. So I think you’re fighting an uphill battle to sell this message to people.
Still, consider yourself preaching to the converted in my case, if I’m accurately paraphrasing your views.
In my eyes, the contentious comparison usually isn’t between a good life and nothingness, it’s between a high-probability good life in the near future and an uncertain probability of a large number of good lives further in the future. Some forms of short term utility maximization can be expected to have negative long-term consequences.
I probably agree with you. But I’d like to point out one possible intuition behind objections to your idea.
Intuition: You are supposed to help people by satisfying their (already fixed and existent) preferences. Not by modifying those preferences to meet reality. Or God forbid, invent those preference ex nihilo.
@sarkology: That’s a succinct statement of my position. Thank you.
How does this discussion extend to person-moments?
How are you distinguishing the “semi-existing” from the couldn’t possibly exist?
Since in most cases the value of an existing person will be higher than non-existing, and at least in current 1st world countries, sufficiently higher to make up for some reduction in value to currently existing people, doesn’t this philosophy imply that male humans of reproductive age are “morally obliged to go around impregnating women” and female humans of reproductive age are morally obliged to go around seeking impregnation?
Do you believe that is a correct conclusion from this philosophy? If not, where is the logical error that is leading me astray?
I’ll admit that I disagree with some of the apparent implications of this philosophy, and perhaps being biased by feeling that if the conclusions are wrong, there must be an error in the philosophy. I also feel it is wrong to value “semi-existing” people’s preferences as equal to currently-existing people’s preferences, but I have trouble identifying exactly what is the right way to reason about those without reasoning incorrectly about currently-existing people’s preferences over their future non-existence.
Um, just so implications don’t get too squicky, I’d like to stipulate that any forcible impregnation or forcible removal of reproductive material will cause sufficient harm as to negate any utility benefit of the people thereby created. I really only want to discuss what consenting people should morally be doing, and just assume any non-consensual actions are addressed elsewhere in the utility calculus, regardless of how we treat potentially existing people’s preferences.
Note that that conclusion doesn’t follow if we use a veil of ignorance argument (see my previous comment). A child who is the result of such a forced impregnation is likely to have a worse than average life, and so to reduce expected utility behind the veil.
I’m previously unfamiliar with veil of ignorance arguments. (I just looked it up on Wikipedia to find out what it means…)
Your claim appears to be that a veil of ignorance argument changes the decision criteria from a better than non-existent life, to a better than average life? Is that the average now, or how does the change in average from the new life effect the decision? Is that a local or global average?
I still don’t think I can agree with the implications. As far as I can see, assuming global averages, it would still be immoral in most cases for a citizen of a first-world country to use birth control during consensual intercourse. (Unless I am mistaken, due to current levels of inequality, even being raised by a poor single mother in a first world country is still better off than the average human on earth.)
There might be an exception for a person who would, with very high probability, have many children at some point in their life, that they could use birth control to optimize the outcomes of those children. However, it still sounds to me like choosing to have no children at all would be usually immoral, as would having less children than you could support in having better-than-average outcomes.
P.S. I later realized my comment could imply use of force, and tried to put a reply that stipulated consenting parties, but it looks like that post has been delayed.
You are right that results are very sensitive to who the reference group is. If the reference group is all people born at the same time, then yes, we probably do have a moral duty in developed countries to have as many children as possible. (I don’t think this is quite the same as contraception being immoral, by not having children now I may be increasing the number of children I can support later on.) However, I don’t think that’s necessarily the correct reference group. In particular, I think most would agree that we have moral duties towards the future, and to capture these duties you need a reference group that includes all people ever. (So behind the veil of ignorance I don’t know when I’m going to be born even.) If this is correct then our imperative to have children now is massively reduced. You might think for example that high population now increases the chance of a cataclysmic event that will hasten humanity’s extinction. Even if this is just a tiny probability it can still massively reduce expected utility behind the veil.
Ow… my head is hurting trying to figure out how I would ever make a reasonable decision about how many children I should be having using a veil-of-ignorance over all time. It also appears that all my possible future children’s preferences get an equal say with my own preferences over which of them should exist or not exist.
There are many conflicting arguments for why additional population could increase (ex. overpopulation) or decrease (ex. more people = more innovation) the odds of future cataclysm. I suspect that even our best current state of knowledge puts such large error bars on these factors that you can’t realistically conclude which is more likely right now. (If you know of a meta-study that puts error bars on this that don’t cross the neutral-point, please provide a link.)
If it is true that it is currently unpredictable whether a given population increase in the near-future will cause a population increase or decrease in the far-future, doesn’t that make such veil-of-ignorance arguments unhelpful?
That, or the conclusion is that I shouldn’t reproduce because I don’t have super-human intelligence… ;)
Well on the links between population and productivity growth, see Jones’s work e.g. http://dx.doi.org/10.1016/S1574-0684(05)01016-6 . The established empirical conclusion is that there are weak scale effects (productivity levels increasing in population levels) not strong scale effects (productivity *growth* increasing in population levels). Weak scale effects means that there’s no incentive to boost population levels now in order to raise (distant) future productivity because there’s no (long-run) history dependence. The contribution of scale effects to future productivity is determined once future population is determined. The knock on links to cataclysm are difficult to assess empirically I guess.
I don’t think it’s a failure really that this theory means the moral status of having children now is uncertain. This coincides with the popular debate on this issue. The theory also tells us what information would enable us to resolve the debate one way or the other. For example, if we encountered an advanced alien civilization that was absolutely committed to preserving intelligent species it encountered (and had been successfully doing so for billions of years) then the moral imperative would switch pretty unambiguously to having lots of children straight away (assuming our reference set only includes humans), since we would know both that 1) our productivity was going to rapidly stabilise at the level of the alien’s technology and 2) we were very unlikely to go extinct. In a less far fetched example, the theory says that neolithic humans had a pretty unambiguous moral imperative to have as many children as possible, since in doing so they reduced the chance of a Neanderthal style extinction event.
I’d like to apologize for this comment, I’m behind on sleep, and apparently enough to impair my thinking ability below the level of this argument.
Really I’ve gone off topic here into what is effectively a generic argument against utilitarian ethics, without realizing it. While I feel that there is something immoral, or at least unfair, about the morality of an action depending on something that is unknowable at the time the action is taken, it really has nothing to do with the topic of how the preferences of non-existent beings should be accounted for.
Pingback: Satisfying preferences by creating them | Meteuphoric
Pingback: Overcoming Bias : Resolving Paradoxes of Intuition
Pingback: Resolving Paradoxes of Intuition | Meteuphoric