Where do I fall if I believe that...
- Trying to add or average different individuals' "happiness" is an invalid move from the get-go (not even the VNM theorem gets you that, and it only holds under counterfactual assumptions).
- Even trying to compare different individuals' happiness is out of bounds in a "pure" sense. It becomes really unreasonable in even a loose, messy, practical sense if you try to go beyond doing it pairwise or take it much beyond "which one is happier".
- There is no way to set a meaningful common origin point between different people.
- There is probably no way to set a "worth living" threshold even for one person (other than asking them, and accepting that the answer will vary widely from one day to the next and depending on what aspects of their situation you happen to highlight).
- It's unlikely that even one person's "happiness" can be meaningfully captured on any one-dimensional scale at all, even at one instant.
- Actual people don't have utility functions, especially if you look at them over time, and there is rarely any value in pretending that they do.
- You do not have the same ethical obligations with respect to each and every person, and furthermore simple proximity and salience can affect your obligations.
- The absolute number of people who exist is allowed to matter in arbitrarily complicated ways.
- Which of two possible worlds is better can depend on how you got there (and leaving that out because it would complicate a quiz is another invalid move).
- There are lots of good reasons other than human happiness to prefer one world over another.
- Your scenario descriptions are therefore meaningless, even though some of the questions feel like they have right answers if you ignore the fact that the whole setup is ill-formed.
- As in all other areas of ethics, expecting consistency within some formal model is unreasonable and of no practical value.
- I can just use whatever feels right, because there is no "true" view, period.
... because I think I'm pretty close to all of those, and, honestly, I think I'm going to mostly ignore population ethics (and all attempts to formalize ethics), even if you don't think I get to.
Sounds like you're describing a more extreme version of the quietist view, in which no outcomes can be ranked against each other. You go even further by saying not only can the described worlds not be compared, but the described worlds aren't even coherent (e.g. there is no such thing as having -7 welfare, regardless of how you define the scale). Although the way to answer the quiz questions is still to say "these outcomes are incomparable".
Interesting. I didn't think much about each question, just answered instinctively. Ended up as averagism with 0 conflicts.
1. Pareto improvement If the very same people all live better lives, that is better.
2. The same number, different people A future of 100 people is worse than a future of 100 entirely different people whose lives all go far better.
3. A against B A is not rankable against B.
4. Adding a life of suffering Adding Nadia with a life of suffering makes the outcome worse.
5. Adding a modest good life Adding Nadia with a modest good life makes the outcome better.
6. Benign addition A+ is better than A — everyone gains, and good new lives are added.
7. Levelling up B is not rankable against A+ — same headcount, more total, more average, fully equal.
8. Repeating the moves Those two verdicts hold identically at every rung of the ladder.
9. A against Z A is not rankable against Z.
10. Adding a wonderful life Adding Nadia with a wonderful life makes the outcome better.
12. Transitivity of better-than “Better than” is transitive.
12. Transitivity of equal-goodness “Exactly as good as” is transitive.
13. Choosing from three None of A, B and Z can be ruled out as best.
- The primary effect of malaria nets is that they prevent the deaths of young children.
- Family planning interventions reduce population growth—making existing family members better off, reducing Malthusian competition, etc.
These are super-easy to reconcile! The claimed contradiction seems to come from an assumption that the main effects of dying of malaria and of contraception use are changes to the rate of population growth: this seems like a pretty weird (natalist?) position - most people care very strongly about both of these things for reasons entirely separate to their effect on overall population growth rates!
For example, let's say that rather than aiming for some population-growth set-point, one is trying to minimise suffering: if you want to raise children but they die of malaria that seems very distressing for your family, and if you don't want children but are obliged to spend the next 18-odd years raising some anyway that also seems pretty distressing for your family. For people that want/love their children nobody seriously believes that the improved financial position from having the children die of malaria makes up for the misery caused by their deaths, and for people who don't want children nobody seriously believes that the social/hedonistic/whatever benefits (if indeed benefits there be..!) of extra children makes up for the misery caused by obliging people to look after children they don't want to look after (and for obliging children to be fully dependent on adults who would prefer them to not exist). Even though they have the opposite-direction effects on population growth, which might-or-might-not be loosely correlated with suffering, malaria nets and contraceptives have same-direction effects on suffering, the thing we actually care about (rather than just a loose potential correlate), and so there's no contradiction here.
(I know I'm only engaging with the first example provided, not the overall argument - but if "malaria vs. contraception" was chosen as #1 because it's the best, cleanest example available, personally I'm inclined to dismiss the argument at this point!)
"Many people hold views that [...] have strange implications that they wouldn't endorse if they thought about it more", "Do your views have counterintuitive implications that most people would dislike?", "There are some ways to reconcile these views, but any attempted reconciliation produces new strange implications", etc.
I think all these claims are entirely true - but not useful. I don't think that there is any fully complete, perfectly-consistent ethical system: they all lead to contradictions, unacceptable edge-cases, etc., and therefore merely showing that some given set of ethical principles leads to this doesn't really tell us anything about the quality of those principles or the people holding them.
Instead, to evaluate sets-of-principles I think we can either start from the assumption that they're all just flawed heuristics and endeavour to show that some set of heuristics seems to work better in more common cases than some other set, or else we can try to do complicated ethical philosophy and build theories with ever-fewer flaws/contradictions than their preceding theories. Perhaps (and I admit I have no idea how realistic this is..) we can even try to do complicated metaethical philosophy and work out whether ethical theories are all incomplete because they relate to incomplete subsets of reality (like Relativity vs. Quantum Physics) or whether they're incomplete for fundamental inescapable reasons (like Russell & Whitehead's Principia Mathematica vs. Gödel's Incompleteness Theorem)?
I would encourage you to watch Hilary Greaves' talk because she explained it better than I could. But the claim isn't that population growth is the thing that matters. Your resolution (as I understand it) is perfectly consistent. The problem is: the way development economists etc. calculate the cost-effectiveness of buying malaria nets is by considering the value of the life-years that would've been lived but aren't. The effect on existing people is real, and it's big in absolute terms, but it's much smaller in relative terms. If you're claiming that the primary beneficiaries of malaria nets are the parents whose children would've died, then that suggests malaria nets aren't very cost-effective compared to a lot of other global health interventions.
So one of two things must be true:
- Against Malaria Foundation much less cost-effective than people say it is
- family planning is much less cost-effective than people say it is
If we take your accounting, then yes they can both be good, but they're not among the best global health interventions anymore.
(Disclaimer; I haven't actually looked at the numbers in a long time so I'm not 100% sure that's true, but IIRC almost all the value of AMF is attributed to the life-years created, not the parents' suffering averted)
Thanks for the reply! I think we're in agreement here about both the lack of contradiction for the given value-system and about the limitations of QALYs as a measure of utility - but I disagree with this section:
So one of two things must be true:
1. Against Malaria Foundation much less cost-effective than people say it is
2. family planning is much less cost-effective than people say it is
So, I'm going to try to rescue the high effectiveness of both of these interventions without having to either introduce a contradiction or give up QALY-sytle measurement!
For example, what if we:
A) Broaden the definition of "quality" in our QALYs (which I gather generally only measures "quality-of-life" from a purely medical perspective) so that we can count such things as "preventing somebody from having to live through the premature death of their child" as an increase in their quality-of-life, and
B) Broaden the comparison of interventions to evaluate not just delta-QALYs for the immediate patient but for any immediate family touched by the intervention?
Plugging-in some arbitrary made-up numbers just to illustrate the effects of broadening the way QALYs are counted (not because the numbers themselves are accurate/necessary):
- Let's say living a perfectly ordinary lifetime counts as 10 quality-of-life * 70 years = 700 lifetime QALYs
- Dying of malaria at age 10: 10 quality-of-life * 10 years = 100 lifetime QALYs
- Watching your child die of malaria when you are age 30: 10 quality-of-life * 30 years + 5 quality-of-life * 40 years = 500 lifetime QALYs
- Being obliged to raise a child you don't want to raise and would have preferred to never have been born, starting at age 20: 10 quality-of-life * 20 years + 3 quality-of-life * 16 years + 10 quality-of-life * 34 years = 588 lifetime QALYs
- Being raised by and fully dependent upon parents who wish you weren't born and would have preferred you to never exist: 3 quality-of-life * 16 years + 6 quality-of-life * 54 years = 372 lifetime QALYs
For a family consisting of two parents and one child, the malaria nets get us 700*3 - (100 + 500*2) = 1000 QALYs and the contraceptives get us 700*3 - (372 + 588*2) = 552 QALYs: that's 1.4 extra [700-QALY] quality-adjusted lives and 0.8 extra quality-adjusted lives respectively. This seems like a lot!
Moreover, since both malaria nets and contraceptives are very cheap, both of these interventions probably work-out at lots of QALYs per unit currency. I think, therefore, that it's perfectly possible for malaria nets and contraceptives to both be very cost-effective ways of improving the world.
A small definitional note: one QALY is defined as a normal life-year (i.e. the thing that you defined as 10 QALYs is actually one QALY).
Measuring EQ-5D-5L utility values in parents who have experienced perinatal death used survey data to estimate that an infant death results in a loss of 1.1 QALYs over the subsequent 10 years, although of course there are a lot of potential issues with this sort of estimate.
I couldn't find if malaria QALY loss estimates incorporate parents' suffering. GiveWell's numbers come from its moral weights spreadsheet, but I couldn't tell how the moral weights were derived.
1I think your argument doesn't actually conflict with Greaves'—it means estimates for malaria prevention cost-effectiveness are wrong twice, but in opposite directions:
- they incorrectly count the life-years lost from the child who died
- they fail to consider the suffering of the parents
(I think Greaves' preferred resolution does count the life-years lost as a bad thing, but she would probably say: if you don't consider it bad to reduce life-years via contraception, then you also shouldn't consider the life-years of the child who died of malaria.)
For your point 2 - fair criticism! If they're not even trying to estimate this, they probably should be.
For your point 1 - I think I am counting the life-years lost from the child who died? If AMF are counting the same way as I, I think we're both doing this?
When I do 700*3 - (100 + 500*2) = 1000 QALYs, I'm doing "QALYs of normal family - QALYs of family with dead child = QALYs lost owing to child death". If we temporarily strip-out the parents from this calculation we get 700 - 100 = 600 QALYs. Isn't this just another way of saying that 600 QALYs, or 6/7ths of a quality-adjusted normal lifetime, are lost owing to child death?
(nb. I'm sticking with the arbitrary 1 ordinary life-year = 10 QALYs figure I plucked from thin air here, not your order-of-magnitude more correct 1 ordinary life-year = 1 QALY, just for consistency)
For your point 1 - I think I am counting the life-years lost from the child who died?
Oh yeah I just didn't read your calculation very carefully, that's my bad.
For the contraceptives case shouldn't the formula be
700*2 - (372 + 588*2) = -148
because only two people exist, not three?
Oh yes, you're right about the formula! Copy-paste error, sorry - I copied the formula from the earlier example and changed the values before solving it, and just missed out changing that value.
What this -148 figure is saying is that the contraceptives provide negative utility in this case because even an unpleasant life, where you spend your formative years fully dependent on parents who would prefer that you were never born, is still better than no life at all.
I'm sure this is true for some people who can overcome the incredible handicap of this upbringing and go on to live overall-happy lives - but personally I think this is pretty rare and that in the general case it'd probably be better to just not be born than to be raised by parents who don't want you, and thus I would argue for lower quality-of-life values than the 3 and 6 I rigorously plucked out of thin air for the childhood and adulthood, respectively.
EDIT: Do contraceptives even prevent people from being born?! I find myself wondering whether people who use contraceptives rigourously might have roughly the same number of children as people who use them haphazardly (or not at all) - perhaps they just have the children at 26 when they actually want to instead of at 16 by accident? If so there wouldn't be negative utility for using them because the contraceptives, instead of stopping children from being born entirely, just convert unwanted children into wanted children...
I was actually thinking a 6/10 quality of life still sounded too low. It seems crazy to me to claim that the average person born to unwilling parents has a net negative life. Even in the United States, 20% of births are unwanted (source). Are you saying those 20% of children go on to have lives containing more suffering than happiness? That seems wild to me.
EDIT: Do contraceptives even prevent people from being born?!
This is also a factor to consider, and empirically it's pretty complicated to work out the answer so I don't know.
I think a great many people have a net negative life, yes, based on how much of our lives we are obliged to spend doing stuff we don't actually want to do, how much awful stuff vs. good stuff happens in the world, etc.
I think it depends a great deal on how you count it, though, and we can get any answer we like by using a different methodology! For example, if we asked everybody every ten minutes to give a yes/no answer on whether the thing they're doing right now is something they actually, genuinely want to be doing, then subtracted all the "no"s from all the "yes"s, I think we'd get a huge negative number. The average person spends 90 minutes per day just sitting in traffic, 7 hours per day at work (most of which involves sitting all-but-motionless whilst staring at a screen, which I do not think represents any part of natural human flourishing), 2 hours doing never-ending mundane household chores, etc. etc. When I look at what the people around me (and the people I encounter online) do and say - nobody seems happy. Everybody seems to me to spend almost all their time doing stuff they actually don't want to do (at every possible level, from childcare duties, to making career-defining promotion-worthy PowerPoint presentations, to whiling away a spare minute scrolling on their portable mini sadness factories phones...), everybody seems angry and upset about stuff all the time, and the world seems full of horrible stuff happening almost everywhere, almost all the time, for almost no legible reasons.
Conversely, I think if you were to approach people on their death-beds and ask them to look back on their lives and rate how happy they were, whether they regret anything, whether they were glad they were born, etc., I think most people would report being pretty happy. Even if you were to accost people at random on the street and ask them to look back at their life so far, I think most people would report that it was mostly happy.
Even in the United States, 20% of births are unwanted (source). Are you saying those 20% of children go on to have lives containing more suffering than happiness?
This statistic* reports unintended pregnancies, not unwanted ones. That's lumping-together an "I don't particularly care to use contraception and if I make a baby then so be it" group, which in my experience includes a fair number of comfortably-off people in stable long-term relationships who are at least partially prepared to raise a child should the need arise, along with an "I actively don't want to make a baby and it'll ruin my whole life if I do" group. I would guess the former group and their children may have life happiness close to average - but I would be astonished if the latter group does!
(* Thanks for the source! It's a pleasure to discuss things with somebody who actually provides citations!)
This statistic* reports unintended pregnancies, not unwanted ones.
From the article (my emphasis):
According to the sheet, 51-plus percent of U.S. pregnancies in 2008–or 3.4 million of 6.6 million–were unintended, which the sheet defines as mistimed (31 percent) or unwanted (20 percent).
Other than that small correction, I'm going to leave it there because I don't have much else I want to say on this topic right now. I'll just add that you are making reasonable points and it's definitely a subject worth thinking about.
11
1How does population ethics interact with resource constraints? Suppose that the world has a fixed amount R of resources, and the
Not 100% sure I understand what you're saying but: To decide what to do given resource constrains, you need to have a model of how those constraints matter (e.g. welfare is proportional to log(resources)), and you also need to have a view of population ethics.
If your goal is to maximize
Very nice, thanks for setting this up!
A way to share your results as an exportable image would make sharing results easier and fun :)
The results are too big to fit cleanly on an image IMO, but the share link at the bottom lets people open a page to see the exact results you got. For example Measure shared their results in a different comment, viewable here: https://mdickens.me/pop-ethics/#a=yrnlrrnynr---y-y-n-&q=r&s=1
Would be nice if the quiz let you look at brief summaries of the other viewpoints at the end. I was curious to retake the quiz with new responses, but am reluctant to do so if you are reading into the results.
Also, my brain places a premium on logical consistency, so after the first few questions, I felt pretty much locked into totalism. If the final question was the lead, I could easily imagine myself getting locked into one of the other camps (e.g., guess I can't compare populations anymore).
I was curious to retake the quiz with new responses, but am reluctant to do so if you are reading into the results.
That's what the name field is for: if you take the quiz multiple times with the same name, I'll just use the first one. (You don't have to use your real name if you don't want to, just use a name.)
On the results page, where it shows a list of all your answers, the question statements are hyperlinks, you can click on them to go back and change your answers if you want to see how it changes the outcome.
I do agree it would be nice to be able to see multiple views, although I'm not sure a good UI for that, and there are a lot of different views. There's a test file with a long list of possible results.
Most people intuitively feel that creating a new life is morally neutral. The standard account of this intuition is the person-affecting view: something can only be good if it's good for a person; a non-existent person isn't worse off for not being born, because there is no such person.
I think these two statements are different. Most people feel that creating a new life is morally neutral because it is (in my view) extremely close to neutral. Neutral enough it's swamped by other factors. But not exactly neutral, and adding people can make the world worse which you could describe as morally wrong.
Also I think there is a big moral difference between an action that effects possible future people and actions that effect current people, most of the paradox goes away with this view.
TLDR: Take the population ethics quiz here: https://mdickens.me/pop-ethics/
Population ethics is an oft-overlooked subfield within ethics. Many people hold views that they don't realize contradict each other, or that have strange implications that they wouldn't endorse if they thought about it more.
Not just that—population ethics is a BIG DEAL. A lot of ethical decisions hinge on how you think about changes in future populations.
I'll give concrete examples below, but reading the examples might change your answers, so you should take the quiz first. It has 13 to 17 short questions. (Questions are skipped when they can't affect the results.)
Cross-posted from my website.
Nobody gets to ignore population ethics
There is no version of population ethics where you can just say, "Whatever feels intuitively right is what's right." Common-sense views on population ethics cannot all be true.
The quiz illustrates that fact. It asks you some questions about how you judge different situations involving varying populations. Then, based on your answers, it figures out:
Answers are always contradictory or counterintuitive or both. The best-case scenario is that your views are consistent, and you bite the bullet on some unpopular conclusion.
If the quiz uncovers a conflict, it will tell you exactly which of your answers are mutually incompatible. Having a conflict is normal. But that's a good prompt to think about how you came to those answers, and how you might resolve the conflict. Did one of your answers come from an intuition that, upon further reflection, you might not agree with? Which intuitions are more reliable than others?
Concrete scenarios where population ethics matters
There are high-stakes real life situations that turn on how we answer population ethics dilemmas.
Example 1: Child mortality vs. family planning
This example comes from Greaves (2015) [1] .
Two of the most well-regarded global poverty interventions are malaria nets and family planning:
The first-order effect of the former is that population goes up; the first-order effect of the latter is that population goes down. On its face, it seems strange to claim that these are both desirable. There are some ways to reconcile these views, but any attempted reconciliation produces new strange implications.
At minimum, we need to change how we estimate the cost-effectiveness of these two interventions. When people calculate the effectiveness of bednets and the like, they typically only consider the effect on life extension, not existing people; when calculating the effectiveness of family planning, they only consider the effect on existing people, ignoring the welfare of the people who would've been born otherwise. Considering both sides makes the interventions look less cost-effective.
I made a related argument in a 2016 post: GiveWell's Charity Recommendations Require Taking a Controversial Stance on Population Ethics. GiveWell recommendations assume that the value of a life saved derives precisely from the extra years of life lived, but also that there is nothing bad about causing someone not to be born (thus preventing them from experiencing 60–80 years of life). I showed that the apparent contradiction in this view is difficult to escape; the most straightforward resolution is that the Against Malaria Foundation is an order of magnitude less cost-effective than GiveWell's estimate suggests.
Example 2: Environmentalism
(I'm going to ignore wild animal welfare because that's a whole other issue. This example is just about how environmental destruction impacts future generations of humans.)
Most people intuitively feel that creating a new life is morally neutral. The standard account of this intuition is the person-affecting view: something can only be good if it's good for a person; a non-existent person isn't worse off for not being born, because there is no such person. However, this view creates some problems. An example: environmentalism.
Suppose a nation has two choices:
The choice has ripple effects into the future. These ripples change which children are conceived—a small change can cause a different sperm to win the race—so the two futures are populated by entirely different people. Therefore, no particular person is harmed by choosing to deplete resources.
This is Derek Parfit's "depletion vs. conservation" case from his analysis of the non-identity problem [2] : if no definite person is harmed by environmental destruction, then person-affecting views have a difficult time explaining how any wrong is done.
These two examples illustrate that there are important real-world situations where we need to consider population ethics.
About the quiz
This project takes inspiration from Philosophy Experiments, which uses interactive philosophy quizzes to illustrate different viewpoints and tease out inconsistencies in your views. They didn't have a quiz for population ethics, so I made my own.
The quiz questions draw on Parfit's mere addition argument (Reasons and Persons, ch. 19 [3] ) and his non-identity problem (ch. 16); Broome's argument that neutrality admits no range (Weighing Lives, ch. 10 [4] ) and his collapsing principle against incomparability and greediness objection to neutrality (ch. 12); and Sen's properties α and β [5] . The results describe positions staked out by Temkin [6] and Rachels [7] (denying transitivity); Tännsjö [8] and Huemer [9] (biting the bullet on the repugnant conclusion); Carlson [10] (against the collapsing principle); among many other positions.
The quiz focuses on ranking outcomes, rather than ranking individuals' choices. Some views hold that the two are different. You might believe that, although world A is better than world B, it would nonetheless be wrong for a particular agent living in world B to bring about world A instead. [11] I didn't address this distinction in the quiz because it would add great complexity.
I'm sure there are also some lesser-known views in population ethics that I overlooked when building the quiz. Similarly, the quiz may flag certain views as contradictory when I failed to see some subtle method to wriggle out of the contradiction.
AI usage disclosure
The initial draft of the quiz was generated wholesale by Claude Opus 4.8, then refined through many rounds of feedback. I vetted a list of population ethics views to ensure that the quiz handled them correctly. [12] Finally, I manually reviewed the text in the final version and revised most of it. The revising process uncovered a few factual errors, but for the most part I only changed wording. [13]
As always, this post itself was written by hand.
Source code
Source code is available on GitHub. Of particular interest is views-review.md, which lists two dozen different views on population ethics and how documents they would answer each quiz question.
Greaves, H. (2015). Repugnant Interventions – EA Global Melbourne 2015. YouTube. ↩︎
Parfit, D. (1984). Reasons and Persons. Oxford University Press. Sections 119–127. ↩︎
Parfit, D. (1984). Reasons and persons. Oxford University Press. ↩︎
Broome, J. (2004). Weighing lives. Oxford University Press. ↩︎
Sen, A. K. (1971). Choice functions and revealed preference. The Review of Economic Studies, 38(3), 307–317. ↩︎
Temkin, L. S. (1987). Intransitivity and the mere addition paradox. Philosophy & Public Affairs, 16(2), 138–187. ↩︎
Rachels, S. (1998). Counterexamples to the transitivity of "better than." Australasian Journal of Philosophy, 76(1), 71–83. ↩︎
Tännsjö, T. (2002). Why we ought to accept the repugnant conclusion. Utilitas, 14(3), 339–359. ↩︎
Huemer, M. (2008). In defence of repugnance. Mind, 117(468), 899–933. ↩︎
Carlson, E. (2004). Broome's argument against value incomparability. Utilitas, 16(2), 220–224. ↩︎
A classic example is the trolley problem. A deontologist might say that, although it would be better for the one person to die instead of the five, it would nonetheless be morally wrong to flip the switch and cause the one to die. ↩︎
I vetted an earlier version, and then had Claude make some changes to the questions. I didn't re-vet after making the changes. For the listed views that allegedly represent the positions of particular philosophers, I did not read the philosophers' works to confirm that their views are represented accurately. The important thing for testing purposes was that the answers describe a view, not whose view it is. ↩︎
Claude generally did quite a good job—we're close to the point where LLMs can create quizzes like this with no human oversight. Claude one-shotted a better UX than any online quiz template I've seen (except maybe GuidedTrack), and Claude correctly understood all of the different population ethics views with only a few minor mistakes. But I didn't much care for its writing style:
Claude generated the quiz using British English. (Maybe because Derek Parfit is British?) When revising the text, I kept the British spellings. ↩︎