This is the second of a series of posts in which I'm trying to build a framework to evaluate aging research. Previous post: A general framework for evaluating aging research. Part 1: reasoning with Longevity Escape Velocity.
Summary
The first part of this post will explore the potential sensitivity of the impact of aging research on two different views of population ethics. Under the person-affecting view of population ethics, in which creating new lives has neutral moral value, it seems that aging research is really valuable. Under the impersonal view, in which creating new lives has a positive moral value, it could be less clear. By looking at demographic trends, and analyzing the motivations for why people have children, it turns out that saving people by hastening the arrival of LEV wouldn't prevent births and could actually increase the average fertility rate of the world. This leads to a counterintuitive result: Aging research could be even more valuable under the impersonal view of population ethics.
In the second part of the post, I'll explore how to reason about moral weights, which could also increase the impact of making LEV come closer if longer lives are valued more than shorter lives for reasons other than QALYs. There are various arguments for why one should prefer some kind of age-discounting or its contrary, but the answer ultimately depends on what a 1000-year-old life and mind looks like and how it is different from the life and mind of a shorter-lived person. Therefore, taking a neutral stance is suggested unless it is believed that the future, if there will be one, is more likely to be better than the present. In that case, the lives of people saved through LEV should count for more.
Population ethics
At first glance, the impact of aging research seems to greatly change depending on if you adopt the impersonal view of population ethics or the person-affecting perspective. In the impersonal view, creating new lives is regarded as good. Assuming that there isn't suffering at the end of life and people get replaced immediately, this view holds no ethical difference between making people live longer and replacing them with new people. Under the person-affecting perspective, however, creating new lives is not valued: only already existing people are valued, and thus how bad it is to die depends on the amount of well-being lost.
MichaelPlant reminded me of this point under my previous post. I gave an answer there but don't think it is sufficient. Therefore, a more accurate analysis of this consideration is warranted here.
It seems like if the person-affecting perspective is adopted, then aging research has enormous value. That is the impact outlined in the previous post.
It seems, though, that aging research could have at least the same value under the impersonal view if elongating healthy life does not mean taking the space of potential newborns. When trying to determine if it would, it's tempting to think about the very far future and start with this question: will humanity use all the resources at its disposal at any given time? Even if humanity will not use all the resources at its disposal, will it still control its population growth in order to maximize well-being? If the answer to one of these two questions is "yes", then it seems like elongating life would prevent births.
Starting with these questions and thinking about the far future is wrong. Reminder: most of the impact of aging research comes from making the date of LEV come closer and saving the people who wouldn't otherwise have hit LEV. If LEV will happen, then it's very probable that it will happen in this century or the next. Therefore, to answer the question "will elongating life prevent births?" we need to account for how society currently works and the current demographic trends.
It seems that the choice of making children is not currently motivated by lack of resources (poverty); on the contrary, the number of children is going down sharply with increased standards of living. This is a trend that is proving true in every part of the world, underdeveloped nations included.
This means that making people live longer in this or the next century is not going to prevent potential births. They will probably happen less and less regardless, and making old people healthy and productive is going to prevent the economic disaster that is looming due to an increasingly aged population, even in underdeveloped countries.
Quite the inverse could prove to be true, though: people with longer lifespans could have more children, simply because they will have much more time to procreate via the childbearing window being extended. Therefore, the fertility rate will probably increase. This consideration could even be the reason why a scenario in which population control will be needed could prove true, although I tend to think that ceiling is very far away in the future, due to technology still having an ample margin of improvement. In case longer lifespans actually increase the world fertility rate, then the impact under the impersonal view of population ethics is the sum of the QALYs saved due to making LEV come closer, plus the QALYs of the newborns of the people saved, who wouldn't otherwise have been born.
Additionally, if longer lives are more valuable than shorter ones for reasons different than the number of QALYs, the neutral view could still value longer lives over perfectly replacing them with shorter lives. This brings us to how to choose moral weights.
Moral weights
An important question that could substantially affect the measure of impact is how to choose moral weights. I think that it's practically impossible to come to a definitive answer due to a lack of empirical information, but I can outline possible ways to reason about the problem.
The central question seems to be: is a 1000-year life intrinsically more, less, or equally as valuable as many shorter lives that sum up to 1000 years?
One argument for why it could be less valuable could be this simple one: it's only one life. I wouldn't find strange if many people would find different shorter lives more important than a single long one because of some kind of intuition regarding a preference for variety or even fairness. This is also supported by the intuition that many people would choose to live a 90% chance of living for a normal human lifespan than a 10% chance of living a 900-year lifespan.
One life, intuitively, is "fresh" only once. Someone may value shorter lives more because they could be, intuitively, more imbued with fresh experiences. Each one goes through infancy, adolescence, and all the other phases of life.
At first glance, the first argument seems weaker: after all, one person is never really the same. The mind changes continuously, and someone could retain very little of themselves living century after century. Would this person experience less novelty? It's possible, unless the future reserves really incredible new experiences and surprises. However, is novelty all there is to consider?
Many lives are, probably, imbued with more novelty, but one long life could mean insight and accrual of knowledge that would be impossible for a single lifespan. Anecdotes of old scientists and luminaries with vast visions of their fields but lacking the sharpness of mind to contribute, especially in hard sciences or mathematics, are common. Each one of them dying is a burnt library of insight and knowledge. Severing their lives at that point means also preventing any future experience resulting from that knowledge. In some sense, it feels like stopping to play when the fun begins, and this could also say something about novelty, which may not be extinguishable very soon. Much longer lifespans could also possibly mean deeper and otherwise impossible-to-experience emotions and states of mind, making longer lives more valuable. This seems obvious if we take, again, the example of luminaries: for a common individual, a normal human lifespan may be not enough to acquire the knowledge of a luminary. Thus, a short life may constitute a hard wall against what can be experienced by most people.
Another intuition that would make one consider a longer life more valuable is this: I think there is a pretty strong case for preferring to have one generation of people living 80 years than multiple generations of children living till the fifth year of age. Therefore, maybe the same intuition could apply for longer lifespans. Are people living till 100 like children if compared to someone living to 1000 years old? The answer to this can't be definitive. I think the answer depends on information we currently don't have: how a 1000-year life and mind looks like and how it is different from the life and mind of a shorter-lived person.
Intuition on how to assign moral weights suggests both issues: If we lean towards valuing longer lives more, we could be overestimating how much more "enlightened" a human mind can become. If we lean towards valuing shorter lives more, we may underestimate the same variable or even commit a mistake akin to scope insensitivity if we don't think about the problem deeply enough.
One consideration that could shift the needle considerably on this is if you deem it probable that the future will be better than the present or if you think, instead, that the far future will be worse. I think that the future is more likely to be of a utopian kind or simply devoid of life than worse than the present, and the probability of future existential risks has to be factored in as a discount of impact, but it's not part of moral weights, so I would tend to ethically value longer lives more than shorter ones for this reason.
However, if you think that the probabilities of the future being better or worse than the present offset each other, then there are good arguments for both methods of applying moral weights, and I would argue to apply neither age-discounting nor the contrary. A neutral stance is probably preferable. That said, different analysts should feel free to think about the problem themselves, and if they believe that one outcome is more likely than the other, they may want to correct these crude estimates.
------------------------------------------------------------
Crossposted to LessWrong
The sense of 'even more valuable' meant here seem to be something like 'more adjusted morally relevant QALYs.' But a total view of population ethics (contrasted with a symmetric person-affecting view) generically massively increases the potential QALYs at stake, and shifts the relative choiceworthiness of different options, so that on the impersonal view life extension is less valuable compared to the alternatives (and thus less of a priority for actual efforts) even if more important in absolute terms:
So I think the counterintuitive result is counterintuitive because it's not asking the right (action-guiding) question, and in action-guiding terms the person-affecting view does much more strongly favor life extension.
I agree.
I used "Counterintuitive", because people tend to think the person-affecting view generates more cost-effectiveness than the impersonal view (see comments under my first post), regardless of how the views affect the comparison with other causes. But yes, adopting the person-affective view seems to make aging research look better in comparison to the other causes you mention, since it negates a lot of their impact. Instead, adopting the impersonal view makes the comparison favour prevention of x-risks that could wipe out literally all of humanity (otherwise aging research looks far better), and probably some interventions regarding non-human animals, also depending on how much you value animals.
Note that this doesn't make aging research worthless to evaluate from an EA perspective. Many people and orgs (eg. Open Philanthropy) donate to more than just two top causes... and aging research seems to be second or third place, probably depending on how much you value non-human animals. Mathematically, it makes sense to differentiate between various top causes in order to reduce risk. Differentiating also makes sense when there are single specific interventions, in a seemingly worse causa area, that may nonetheless be more cost-effective than available interventions in a cause-area that overall looks better, which includes cases in which the more cost-effective interventions in the top cause-areas are funded, or if there are particularly cost-effective interventions in the seemingly worse cause-area.
A thought about how LEV relates to the economics of supporting an aged population. If LEV makes older people healthy and productive this also benefits younger people, as less of their resources (either taken as government taxes or spent in person with relatives) are required to support the elderly. So I think this would further supports LEV the benefits of (under the person-affecting view) for younger people and goes against the idea that it's better to package life into shorter units.
One question I have about these discussions is that I'd read some arguments back when I took a class on environmental science that humanity was near the earth's carrying capacity, meaning there is not capacity on earth for a much larger population (and capacity is quite likely smaller). This term "carrying capacity" seems like a sketchy one that tacitly packs in normative and positive judgments, so I don't endorse it, but is there a chance that something like this is true, and so lengthening life will reduce the probability of others being born because it raises the probability of environmental problems that lower the sustainable population?
This is true, but the carrying capacity increases as technology improves. This plus the fact that birthrates are under the replacement rate in the developed world and going down pretty much everywhere should make us think we will not be in a malthusian situation when LEV arrives.
Good post. Some further considerations on the total view side of things (mostly culled from a very old working paper I have here where I suggest life extension may be bad - but N.B. besides its age and a few errors, my overall view is now tentatively pro rather than tentatively con).
0. LEV or not seems to be a distraction. The population ethics concerns don't really change much either way if the offer on the table is LEV or merely 'L' (e.g. there's a new drug which guarantees lifespan to 150 but no more).
1. As the contours of your argument imply, I think the core ethical issue on totalist-y lights would be whether there is a 'packaging constraint' on how one should allocate available lifetime to persons (e.g. better 1 800 year life versus 10 80 year lives, or vice versa), versus a broad cloud of empirical considerations and second order effects (although I think these probably dominate the calculus).
2. I don't buy the story that life extension can be a free lunch. If it is better to 'package' lifespan into 80 year chunks versus millenia-sized chunks, whether or not to pursue this will have great impact across the future, so any initial 'free benefit' will be probably outweighed by ongoing misallocation across the future. (I suppose the story could be 'LEV, even if bad, is inevitable, and doing it sooner at least gets a bigger free lunch - but it seems in such a world there bigger scale problems to target).
3. On pure aggregation, the key seems to be whether lifespan has accelerating or diminishing marginal returns. As you say, intuitive survey by time-tradeoff gives conflicting recommendations: most would be averse to gambles like "Would you rather 5% chance of 2000 years (and 95% of dying right now) versus keeping your life expectancy?", yet we'd also be averse to 'Logan's run' (or Logan's sprint) cases of splitting 80 year lives into 16 5-year lives (or, indeed, millions of 2 minute ones).
3.1 One natural reply to defuse 'Logan's run' type reductios is to suggest it is confounded with human development. One might say our childhood and adolescence is in part an investment to enjoy the greater goods of adulthood. So perhaps we would take lifespan to have accelerating returns up commensurate to this, but maybe not for the interval of 20-ish to infinity (so if the returns diminish, there will usually be a break-even point whereby the 'investment cost' is matched by the diminishing returns loss, so making the ideal tiling of lives across time not 'as long as possible'.
(We should probably be pretty surprised if the morally 'optimal' lifespan just-so-happened to match our actual lifespan which emerged from a mix of contingent biological facts. Of course, it could be the 'optimal' lifespan is shorter, not larger, than the one we can typically expect.)
3.2 There's a natural consideration for diminishing returns on the idea that people may naturally prioritise the best things to do with their life first, and so extending their lives gives them opportunity (borrowing a bit from Bernard Williams) to engage in further projects which, although good, are not as good as those they prioritised before then. So packaging into smaller chunks offers the ability for the population over the time to complete more 'most valuable' projects.
3.3 On the other, there's a murkier issue about maybe having a much longer life 'unlocks' opportunities which are better than those shorter lives can access. In the same way 'living each day as your last' when taken literally is terrible advice (many things people want to do take much longer than a day to accomplish), perhaps (say) observing changes over cosmological or geological timescales are much experiences than what one can do in decades. This looks fairly speculative/weak to me.
What seems more persuasive on the 'increasing marginal returns' side is the idea of positive interaction terms between experience moments. Some good things could be even better if they resonate with other previous moments, and so a longer prior life seems to provide further opportunity for this (e.g. insofar as 'watching the grandchildren grow up' is joyful, a longer life better ensures this occurs, among many other examples).
4 Egalitarianism, 'justicy'-considerations, or prioritarianism will generally push towards packaging in shorter blocks rather than longer ones: the one which best gets around tricky different number cases is prioritarianism. Insofar as you are sympathetic to these views, these will seem to push against life extension.
4.1: I'm pretty sympathetic to Parfitian/deflationary accounts of personal identity, which would take the wind out of the sales of this line of argument (as there isn't much remaining sense of a given person being better or worse off than another, nor of an index to which there's a 'you' that accrues person moments which may have diminishing returns). Such a view also takes the wind out of the sails of a pro life extension case (as we should be relatively indifferent to whether future moments are linked to our present ones or otherwise), although there might be second order considerations (beyond those mentioned above, if most experience moments simply prefer to be linked up to more future ones, this is a pro tanto consideration in favour).
5 It seems the second order impacts are best distinguished from the 'pure axiological' issue above. It could be that very long lives are an imperfect allocation, but still best all-things considered if (for example) it allows people to develop much greater skill and ability and (say) produce works of even greater artistic genius. A challenge to trying to disentangle this is plausible scenarios which offer (radical) life extension likely involve other radical changes to the human condition: maybe we can also enhance ourselves in various ways too (and maybe these aren't seperable, so maybe the moral cost we pay for improperly long lives is a price worth paying for the other benefits).
5.1 If we separate these and imagine some naive 'eternal (or extended) youth' scenario (e.g. people essentially like themselves, with a period of morbidity similar to what we'd expect now, but their period of excellent health extended by a long time), I'd agree this leans positive. Beyond skill building benefits, I'd speculate longer lives probably prompt less short-sightedness in policy and decision making.
Re 3/3.1: When discussing the marginal returns on a human life, a quantitative way of modelling human capability could be as the product of sigmoidal curves with positive and negative slopes to represent the scaling up of capability during development and scaling down of capability during natural aging. As long as aging doesn't kick in before development is finished then there is a plateau phase during which a person can perform at maximum capability and should produce constant returns on extra years in this phase.
Treating treating human capability as a single curve might be too simplistic. One could further break this down to intellectual and physical capability and intrinsic and extrinsic factors:
-Physical capability is simplest as humans probably reach peak intrinsic physical capability around 20 (sharp increase) and start to decline after 40 (slow decline). I'm not sure there are extrinsic factors related to physical capability that will change as a function of a person's life span.
-Intrinsic intellectual capability could probably continue to scale up for a long time with a slow increase (some luminaries may currently get close to peak intellectual capacity, but I suspect that most people alive at the moment don't) and this does not necessarily decline much during aging unless somebody gets an age related neurological disorder (which can cause a very sharp decline). While some might argue that people will keep increasing intellectual capability with age, I'd argue that there probably is an upper limit to intrinsic intellectual capability given the brain's capacity to store and process information (although neurotechnology may extend this). However, extrinsic intellectual factors like professional network size, strength, and value generally do continue to increase over time and could be modelled as a curve with a slow increase; while social network size currently tends to decline in old age this seems to be related to declining physical capability (reduced stamina limiting ability to socialize and forcing retirement) and so improving physical health during old ago may also prevent decline in some extrinsic intelectual areas.
Productivity could then be judged as weighted sums and/or products of intrinsic and extrinsic intellectual and physical capability. The weighting will probably depend on the state of the society an individual lives and would change over time - subsistence farming would weight physical capability strongly, developed society initially favoured intrinsic intellectual capability but increased digital connectivity is increasing the value of extrinsic intellectual factors.
The reason I focus on a model composed of weighted sums/products of sigmoidal curves of positive/negative slopes is that these can actually create fairly interesting results. The sum or the product of two sigmoidal curves with opposing slopes will be something like a bell curve (although it can be flat topped and have asymmetric sides), which probably agrees quite well with how people would judge the productivity of a current human life-span. However, having three sigmoid curves with the result depending on the product of two of them can create a local maxima before a later plateau, which could be used to represent an early peak in productivity due to physical capacity that will later be exceeded by intellectual capability (see this figure for an example I used of such a model https://www.nature.com/articles/srep02614/figures/5 ). Also, sigmoidal curves are quite good at describing many biological processes.
In summary, the point I'm getting at is there could be a good biological/psychological framework to discount life years based on both development and aging.
*Note that I don't have much experience in population ethics and am implicitly equating productivity to value and this may not be a good ethical framework (although I assume it will probably be agreeable to economists!).
Re 3.2: People also often do riskier things earlier in their lives. You don't see many 50 year old startup founders, maybe because they a more likely to need guaranteed income to to support their kids and/or for retirement savings. But their greater knowledge and connections may give them a greater chance of success at high-risk/high-reward type endeavours, and so LEV may allow people to undertake such promising activities later in life when they are better prepared for them.
I think it's an interesting cause area (upvoted for investigating something new), though I have three important quibbles with this analysis (in ascending order of importance):
1) The person-affecting (PA) view doesn't make this a slam-dunk. PAness doesn't signify that death in itself has negative value, so given your assumption 'that there isn't suffering at the end of life and people get replaced immediately', on the base PA view, increasing lifespans wouldn't in itself generate value. No doubt there are flavours of PA that would claim death *does* have disvalue, but those would need to be argued for separately.
Obviously there often *is* profound suffering at the end of life, which IMO is a much stronger argument for longevity research - on both PA and totalising views. Though I would also be very wary of writing articles arguing on those grounds, since most people very sensibly try to come to terms with the process of ageing to reduce its subjective harm to them, and undoing that for the sake of moving LEV forward a few years might cause more psychological harm than it prevented.
2) My impression is that the PA view is held by a fairly small minority of EAs and consequentialist moral philosophers (for advocates of nonconsequentialist moral views, I'm not sure the question would even make sense - and it would make a lot less sense to argue for longevity research based on its consequences), and if so, treating it as having equal evidential weight as totalising views is misleading.
It's obviously too large a topic to give much of an inside view on here, but if your view of ethics is basically monist (as opposed to dualist - ie queer-sort-of-moral-fact-ist) I don't think there's any convincing way you could map real-world processes onto a PA view, such that the PA view would make any sense. There's too much vagary about what would qualify as the 'same' or a 'different' person, and no scientific basis for drawing lines in one place rather than another (and hence, none for drawing any lines at all).
3) 'Reminder: most of the impact of aging research comes from making the date of LEV come closer and saving the people who wouldn't otherwise have hit LEV.'
This is almost entirely wrong. Unless we a) wipe ourselves out shortly after hitting it (which would be an odd notion of longevity), or b) reach it within the lifespans of most existing people *and* take a death-averse-PA view, the vast majority of LEV's impact of it will come on the ripple effect on the far future, and the vast majority of its expected impact will be our best guess as to that.
EAs tend to give near-term poverty/animal welfare causes a pass on that estimation, perhaps due to some PA intuitions, perhaps because they're doing good and (almost) immediate work, which if nothing else gives them a good baseline for comparison, perhaps because the immediate measurable value might be as good a proxy as any for far-future expectation in the absence of good alternative ways to think about the latter (and plenty of people would argue that these are all wrong, and hence that we should focus more directly on the far future. But I doubt many of the people who disagree with *them* would claim on reflection that 'most of the impact of poverty reduction comes from the individuals you've pulled out of poverty').
Longevity research doesn't really share these properties, though, and certainly doesn't have them to the same degree, so it's unlikely to have the same intuitive appeal, in which case it's hard to argue that it *should*. Figuring out the short-term effects is probably the best first step towards doing this, but we shouldn't confuse it with the end goal.
The PA view doesn't need to assign disvalue to death to make increasing lifespans valuable. It just needs to assign to death a smaller value than being alive. Unless you literally don't care if people live or die or you think that dying is better than living, my argument holds.
If we make LEV nearer we don't increase the distress anti-aging therapies will cause to people at first. We just anticipate the distress.
I guess so. I used the same style being used in the introductory articles to EA, which are pretty neutral, although they recognise the neutral view as probably superior. This doesn't matter though, since, as I wrote, impact under the neutral view is actually bigger.
Financing aging research has only the effect of hastening it, so moving the date of LEV closer. The ripple effect that defeating aging would cause on the far future would remain the same. People living 5000 years from now wouldn't care if we hit LEV now or in 2040. So this isn't even a measure of impact.
If you are curious, Sarah Constantin recently wrote an analysis using the shorter term effects of aging research as a measure of impact. This one. Also, my next post is exactly on the shorter term impact. I think it'll be published in a couple of weeks. It will cover DALYs averted at the end of life, impact on life satisfaction, the economic and societal benefits, impact on non-human animals.
It depends how you interpret PA. I don't think there is a standard view - it could be 'maximise the aggregate lifetime utility of everyone currently existing', in which case what you say would be true, or 'maximise the happiness of everyone currently existing while they continue to do so', which I think would turn out to be a form of averaging utilitarianism, and on which what you say would be false.
Yes, but this was a comment about the desirability of public advocacy of longevity therapies rather than the desirability of longevity therapies themselves. It's quite plausible that the latter is desirable and the former undesirable - perhaps enough so to outweigh the latter.
Your argument was that it's bigger subject to its not reducing the birthrate and adding net population in the near future is good in the long run. Both are claims for which I think there's a reasonable case, neither are claims that seem to have .75 probability (I would go lower for at least the second one, but YMMV). With a .44+ probability that one assumption is false, I think it matters a lot.
Again this is totally wrong. Technologies don't just come along and make some predetermined set of changes then leave the world otherwise unchanged - they have hugely divergent effects based on the culture of the time and countless other factors. You might as well argue that if humanity hadn't developed the atomic bomb until last year, the world would look identical to today's except that Japan would have two fewer cities (and that in a few years, after they'd been rebuilt, it would look identical again).
Looking forward to it :)
Good points, although I'm not sure who would hold averaging utilitarianism. But yes, in this case prolonging life wouldn't matter.
I doubt that the damages of public advocacy would outweigh the good. Only if advocacy is really good at convincing people of the possibility of bringing aging under medical control, the large-scale distress you mention could happen. But then aging would become an issue under the eyes of everyone and funding would immediately spike up, along with policies to accelerate the process. If this happens, the supposed psychological distress would be a rounding error if compared even only with additional DALYs prevented at the end of life. Otherwise, if advocacy manages to convince people of the possibility of putting aging under medical but doesn't bring additional money and talent in research, then yes the psychological damage would probably outweigh the positive impact. But is this a possibility? I don't think it's possible to convince a large fraction of the population and at the same time not cause resources to pour in the field. Then you could argue that research could be so ineffective that pouring resources into it wouldn't accelerate anything. But I think this has a very low probability. Note also that in expectation even a very small hastening of the field would outweigh psychological distress.
At worst the PA view and the impersonal view have the same effect, so "it matters a lot" seems exaggerated to me. A totally unrelated idea would be introducing a discounting of impact because of these considerations, but it still wouldn't be advisable using expected value.
I think you are right here, but I still don't think most of the impact would come from the ripple effects that hastening aging research would have on the far future. We don't even know if the effects will be good or bad. In my view they would be probably just cultural and neutral cost-effectiveness wise.