Ethics is about decisions, and most of the discussions about the "difficulty" of infinite ethics only work if you ignore that.
(And so it's particularly frustrating that they didn't bother addressing our work pointing out why they are wrong: https://philpapers.org/rec/MANWIT-6 )
Did not finish reading either yet, but is this not addressed (without explicitly addressing your work) early on in the article?
As I read their reasoning, even if the by far most likely outcome is that your conclusion holds in practice, there is a non-zero probability that assumptions are wrong in a way that allows for infinite causality, and therefore (by assuming their infinite-fanatical stance), attempting that is still a sane conclusion.
More fundamentally, they are reasoning within the infinite set of imaginable universes whereas you reason within our current one and that current consensus of physical limitations hold. Your scope is "only" our morally relevant universe. Different fundamental assumptions yield fundamentally different conclusions.
If you want to use the concept of "probability" to deal with reality, you better decide first if the reality offers infinite payoff/risk. If it does, your tool (the naive probability) is inadequate and quite easily broken, as you just have demonstrated.
The tool will sway to 100% and back to 0% on like a broken compass, depending on which infinities you thought about this very minute.
Good point - it should follow that given assumption that with infinitesimal probability of infinite payoff, there is also a corresponding non-zero probability of infinite risk.
Overall I think OP is mostly about them struggling to reduce hypothetical outcomes involving infinities to real numbers in order to be able to rank them. For example:
> Agent-neutrality: If there is a welfare-preserving bijection from the agents in w1 to the agents in w2, then w1 and w2 are equally good.
> By “welfare-preserving bijection,” I mean a mapping that pairs each agent in w1 with a single agent in w2, and each agent in w2 with a single agent in w1, such that both members of each pair have the same welfare level.
They then proceed to compare a w1 with a 1/0 of 1:1 vs w2 with a 1:3 ratio, "skipping" 2/3 of 0s in w2, claiming a contradiction. So they're really doing a injective non-surjective mapping, not a bijective one, and them being infinite doesn't really absolve that. I think they could get some insight from algorithmic analysis and limits. Just like (probably sloppy notation here) O(0.25n) is faster than O(0.5n) even if both are O(0.5n).
> They then proceed to compare a w1 with a 1/0 of 1:1 vs w2 with a 1:3 ratio, "skipping" 2/3 of 0s in w2, claiming a contradiction. So they're really doing a injective non-surjective mapping, not a bijective one.
There is in fact a bijection between indices that maps the sequences to each other -
I can only reason about infinitely many universes if they're sufficiently similar to my own universe. But in that case I might as well just forget about the whole infinity thing since they're all independent and average out. If something is worth doing here on Earth then it's worth doing.
Also the whole "acasual" thing never made sense to me. What I do doesn't influence anything in another universe by definition, so other universes shouldn't influence my decision making at all.
This is a really interesting topic but I'm offput by the verbosity of this essay. For example:
> And whatever our credences here, we should be clear-eyed about the fact that helping or harming an infinite number of people would be an extremely big deal. Saving a hundred lives, for example, is a deeply significant act. But saving a thousand lives is even more so; a million, even more so; and so on. For any finite number of lives, though, saving an infinite number would save more than that. So saving an infinite number of lives matters at least as much as saving any finite number – and very plausibly, it matters more (see Beckstead and Thomas (2021) for more).
Is it really necessary to have a whole paragraph explaining that saving infinite lives would be a big deal? I don't think a single person would believe otherwise, which means this paragraph is zero entropy. Most philosophical essays seem to have a similar issue.
> I don't think a single person would believe otherwise,
It depends. Saving "a hundred" or a "a thousand" lives seems "doable" by a normal person as a result of a conscious, "moral" act, saving "millions" devolves into messianism, and as such it's not the job of essays or of simple, logical human thinking to talk/write about it. I know that the author doesn't insist on that aspect but in real life quantitate differences turn into qualitative differences most of the time.
With growing populations, I'm not sure trying to save a million lives is THAT special anymore. Definitely still fairly awesome, but not Messiah level awesome.
I think eg. President Zelensky (the buck stops with him for >40M souls) would be very surprised to learn of his sudden deification.
There are actually probably quite a number of rather less spectacular (eg. bureaucratic, military, logistics, sysop, ...) jobs where you might routinely touch 1M+ people these days, in some narrow but nevertheless (up to) life-critical ways.
This does mean one might need to take ethics into account in situations that seem far removed from the coal face, typically whilst pushing dusty papers around; lest you end up creating catch 22's for the people in the trenches.
Everyone has stories about things like unethical corporate behavior. At that kind of scale, it makes more sense to build systems (governments/corporations/foundations/smart mobs/...) that are somehow intrinsically ethical or at very least inspire ethical behavior in their members.
- I feel some analogous thinking with computational complexity. An action may have broad repercussions, but in big O thinking, as we approach infinity, only those factors with the most significant impact matter. Considering infinities is a useful lens when trying to reduce complexity to the most significant factors.
- That kind of analysis is typically only useful if you have the entire algorithm and all its variables mapped out. A big-O approach to the travelling salesman problem wouldn't be accurate if you didn't consider the travelling part. You'd say hey, only 1 destination is added at a time, so the impact of each new destination is linear, which is wildly incorrect.
- As long as a subject matter is subjective and people get to decide which variables belong and which can be ignored, and what impact a variable may have, the analysis will be incorrect, most likely wildly incorrect. Is there a universal method for deciding which factors are important?
"Should we bomb our neighbours and steal their chocolate, or buy their chocolate instead of making our own? Well, our main export is chocolate, so they are taking bread off our citizens tables. Infinite bombs = infinite bread." Obviously a few factors are missing - but which factors?
Obviously there's limits to the technique but I think it's an interesting lens to view ethical concerns through, but I don't see how (would like to better understand if?) it provides any insight into universal ethics or if there is such a thing because at the core of it all is a person who gets to decide what they think is and isn't important to factor into the calculation. Or maybe that's the point I missed.
This formidable essay is beautiful and pure Kierkegaardian tragedy at
the same time.
I feel how much the author wants to extend kindness to the infinite.
Yet we live in a world so atomised and self-objectified that most of
us fail to apply ethics to a few loved ones, to even one person,
ourselves. Real ethics is finite, embodied, pragmatic, messy. It's the
shit, piss, blood and mud of battlefields, caring for the old and
dying, giving birth, honouring or betraying lovers, paying or cheating
the tax man... and 80 years on this tiny rock, if we're lucky, is but
a moment.
No wonder bright minds wish to escape, to be Luftmenchen soaring in
the stratosphere of number theoretical grandiosity and the comfort of
speculative theoretical physics. This is why we invented religions,
and will continue to reinvent them in ever shifting guises.
If we care about ethics, love your neighbour as yourself. Not a
hypothetical neighbour on planet Zarkphobble in the year 20123894, but
that smelly homeless guy in a pool of vomit outside Costco you passed
this morning. And for the rest of us, let's focus on doing less harm
with the technology we create.
> This is why we invented religions, and will continue to reinvent them in ever shifting guises.
Something off about this statement, though. Like out of place in some otherwise wise observations.
Man has our escapism but it isn't at all uniquely circumscribed by "religion".
Unless you define, circularly, "religion" to be man's escapism.
I'm trying to avoid some apologetics over some theology or something. For myself, I'm an apostate. Most religions people would call me an atheist, but I sense that most atheists would reject my membership application.
But what I see in essays like this is a limited human attempting the impossible task of reconstructing some meaning to existence.
The task being impossible b/c we can no more reconstruct meaning than we could reconstruct our, say, biological selves. That was the job of eons of evolutionary process.
Which is what "religion" hosts, encodes... along with crufty theology, escapism, etc.
What you seem to be pitching is this "Just be good" philosophy that for some reason I keep seeing show up on T-shirts the past couple of years. But this is not a substantive philosophy.
When I left religion in my early 20s, in retrospect I feel I threw the baby out with the bathwater not realizing that there was deep, evolved meaning and framework amid the mounds of garbage. That remained even when I thought I was rejecting all of it, the escapism, dogma, fear, etc... In other words I couldn't discard whatever was encoded in myself.
But I worry now I'm not giving my kids what I'd gotten and took for granted. Worry now that a whole generation is not getting it. That we're going to be left with "Just be good" - and otherwise these desperate, searching, doomed essays from people who realize that "Just be good" is not good enough but who have no heritage from the past that hosts pretty much all we could expect to have.
>Something off about this statement, though. Like out of place in some otherwise wise observations. Man has our escapism but it isn't at all uniquely circumscribed by "religion". Unless you define, circularly, "religion" to be man's escapism.
Exactly. Few things have been more pragmatic, and with more practical impact in history (both good and bad), than religion. The "fairy tale" part of it is just a distraction to the actual symbolic content.
Sorry if I sounded dismissive in using the word "escapism". I did not
mean it in a bad way. We all need an escape, and will get it one day.
Years of suffering dogma as a secular humanist returned me to some
kind of faith. Against the grain I take my kid to church. She likes
the songs. I haven't given up on religion. There's no "just" in being
good, We all need help with that.
This was a hard read, and admittedly I didn’t finish it. I feel a little dumb.
I’m hoping someone smarter than myself can momentarily stoop down to my level and clear things up.
I wonder, must one strive to understand the mathematics of ethics to make ethical decisions? Am I able to optimise my decisions for universal wellbeing without being able to mathematically reason upon it?
Something I’m equally ill-equipped at is being present - aware of my real-time inclinations, motivations, and subsequent actions. As a result, I often make decisions which oppose ethical optimisation (or my understanding of it)
If I were to choose between the ability to mathematically reason on ethical matters, or to have the awareness to actually make (potentially misguided) ethical decisions, I’d choose the latter every time.
Near the start a discussion of if you should go on holiday, or save a life. Except there will be an infinite number of slightly different universes where, by going on holiday, you are given the opportunity to save a life. There will also be an infinite number of universes where your donation ends up causing the end of the world.
Once you consider every possible world and every possible outcome, happening infinitely many times, I would draw a different conclusion -- unless you have the power to effect the entire "infinite-verse", then anything you do becomes unimportant, as infinite+finite=infinite, so your contributions are lost in the wash.
This is where measure theory comes in. In spite of the fact that the interval [0, 1] is uncountably infinite, as is every possible subinterval, we can still coherently talk about some event spaces being more probable than others.
This also applies to a single universe where an action might have many unintended consequences, possibly snowballing into exactly the opposite from what you would have wanted. In this sense, each of your decisions are important and you are responsible for everything.
So in such a dynamical system, how do you then act? The philosophically Western answer to this is in part through Logos. You treat the spoken and written word as divine, through truthful and authentic expression of the individual, as clearly as possible. The consequences of how this unfolds is outside your immediate control, as it should be.
Love has an important role here, but Love without Truth seems to be the prevailing attitude online, which is declared as a personal virtue and experienced as righteous indignation. This is to me what others perceive as the fall of the West: the prioritization of Love over Truth.
I can recommend chapter XI What’s the most bullet-biting hedonistic utilitarian response we can think of for a relatively easy read. Chapters II to X would probably require much more upfront investment than I'm personally prepared to make - the TLDR was sufficient.
I couldn't understand why "If you donate, this is strong evidence that they all donate, too" was true. Wouldn't there be infinite choices made otherwise?
The "strength" of the evidence is arguable, but it is definitely evidence. If the risk calculus you did ended with you choosing donating, then you know that the probability is non-zero that each incarnation of you would choose to donate. But not "strong" evidence -- this is flipping a coin once, getting heads, and saying you have strong evidence that the coin is biased towards heads. You have evidence, but I'm not going to stake much on it.
As for the infinite aspect, infinity doesn't mean uniform - there are an infinite number of odd numbers but none of them are even.
While the essay eventually explores some interesting topics, the opening statements claiming there are explanations from physics that connect these infinities to our actual universe were fairly ameteurish and misleading.
The many worlds interpretation of quantum mechanics, for instance, provides clear math on how to reason about quantities as the universe splits into more universes: we integrate after multiplying by the probability of each branch. No credible physical theory is proposing we double the amount of mass and energy in the universe whenever we measure the spin of an electron.
Calling that split universe "two universes" instead of "two half-universes" is fine for a casual explanation, but if you want to do correct math with this result, you need to start with the correct math too! Once you make that adjustment, all this talk in about "finite benefits don't matter compared to infinite benefits" can be translated as "benefits with probability zero don't matter compared to benefits with non-zero probability". Much more palatable.
There are still interesting conversations about benefits which could last for infinite time. But there, the crux is in the "could". These thought experiments always bake in assumptions at dozens of levels which say "this factor in my model will last for infinite time". When you consider in our ignorance about what the future will hold, I've yet to see a convincing example where we could truly consider a finite action as having infinite utility. I'm certainly open to it, I just haven't seen it.
27 comments
[ 4.0 ms ] story [ 84.2 ms ] thread(And so it's particularly frustrating that they didn't bother addressing our work pointing out why they are wrong: https://philpapers.org/rec/MANWIT-6 )
As I read their reasoning, even if the by far most likely outcome is that your conclusion holds in practice, there is a non-zero probability that assumptions are wrong in a way that allows for infinite causality, and therefore (by assuming their infinite-fanatical stance), attempting that is still a sane conclusion.
More fundamentally, they are reasoning within the infinite set of imaginable universes whereas you reason within our current one and that current consensus of physical limitations hold. Your scope is "only" our morally relevant universe. Different fundamental assumptions yield fundamentally different conclusions.
Does that make sense?
The tool will sway to 100% and back to 0% on like a broken compass, depending on which infinities you thought about this very minute.
Overall I think OP is mostly about them struggling to reduce hypothetical outcomes involving infinities to real numbers in order to be able to rank them. For example:
> Agent-neutrality: If there is a welfare-preserving bijection from the agents in w1 to the agents in w2, then w1 and w2 are equally good. > By “welfare-preserving bijection,” I mean a mapping that pairs each agent in w1 with a single agent in w2, and each agent in w2 with a single agent in w1, such that both members of each pair have the same welfare level.
They then proceed to compare a w1 with a 1/0 of 1:1 vs w2 with a 1:3 ratio, "skipping" 2/3 of 0s in w2, claiming a contradiction. So they're really doing a injective non-surjective mapping, not a bijective one, and them being infinite doesn't really absolve that. I think they could get some insight from algorithmic analysis and limits. Just like (probably sloppy notation here) O(0.25n) is faster than O(0.5n) even if both are O(0.5n).
There is in fact a bijection between indices that maps the sequences to each other -
bijective map such that w2(f(i))=w1(i) :Also the whole "acasual" thing never made sense to me. What I do doesn't influence anything in another universe by definition, so other universes shouldn't influence my decision making at all.
> And whatever our credences here, we should be clear-eyed about the fact that helping or harming an infinite number of people would be an extremely big deal. Saving a hundred lives, for example, is a deeply significant act. But saving a thousand lives is even more so; a million, even more so; and so on. For any finite number of lives, though, saving an infinite number would save more than that. So saving an infinite number of lives matters at least as much as saving any finite number – and very plausibly, it matters more (see Beckstead and Thomas (2021) for more).
Is it really necessary to have a whole paragraph explaining that saving infinite lives would be a big deal? I don't think a single person would believe otherwise, which means this paragraph is zero entropy. Most philosophical essays seem to have a similar issue.
It depends. Saving "a hundred" or a "a thousand" lives seems "doable" by a normal person as a result of a conscious, "moral" act, saving "millions" devolves into messianism, and as such it's not the job of essays or of simple, logical human thinking to talk/write about it. I know that the author doesn't insist on that aspect but in real life quantitate differences turn into qualitative differences most of the time.
I think eg. President Zelensky (the buck stops with him for >40M souls) would be very surprised to learn of his sudden deification.
There are actually probably quite a number of rather less spectacular (eg. bureaucratic, military, logistics, sysop, ...) jobs where you might routinely touch 1M+ people these days, in some narrow but nevertheless (up to) life-critical ways.
This does mean one might need to take ethics into account in situations that seem far removed from the coal face, typically whilst pushing dusty papers around; lest you end up creating catch 22's for the people in the trenches.
Everyone has stories about things like unethical corporate behavior. At that kind of scale, it makes more sense to build systems (governments/corporations/foundations/smart mobs/...) that are somehow intrinsically ethical or at very least inspire ethical behavior in their members.
- I feel some analogous thinking with computational complexity. An action may have broad repercussions, but in big O thinking, as we approach infinity, only those factors with the most significant impact matter. Considering infinities is a useful lens when trying to reduce complexity to the most significant factors.
- That kind of analysis is typically only useful if you have the entire algorithm and all its variables mapped out. A big-O approach to the travelling salesman problem wouldn't be accurate if you didn't consider the travelling part. You'd say hey, only 1 destination is added at a time, so the impact of each new destination is linear, which is wildly incorrect.
- As long as a subject matter is subjective and people get to decide which variables belong and which can be ignored, and what impact a variable may have, the analysis will be incorrect, most likely wildly incorrect. Is there a universal method for deciding which factors are important?
"Should we bomb our neighbours and steal their chocolate, or buy their chocolate instead of making our own? Well, our main export is chocolate, so they are taking bread off our citizens tables. Infinite bombs = infinite bread." Obviously a few factors are missing - but which factors?
Obviously there's limits to the technique but I think it's an interesting lens to view ethical concerns through, but I don't see how (would like to better understand if?) it provides any insight into universal ethics or if there is such a thing because at the core of it all is a person who gets to decide what they think is and isn't important to factor into the calculation. Or maybe that's the point I missed.
Interesting in any case.
I feel how much the author wants to extend kindness to the infinite. Yet we live in a world so atomised and self-objectified that most of us fail to apply ethics to a few loved ones, to even one person, ourselves. Real ethics is finite, embodied, pragmatic, messy. It's the shit, piss, blood and mud of battlefields, caring for the old and dying, giving birth, honouring or betraying lovers, paying or cheating the tax man... and 80 years on this tiny rock, if we're lucky, is but a moment.
No wonder bright minds wish to escape, to be Luftmenchen soaring in the stratosphere of number theoretical grandiosity and the comfort of speculative theoretical physics. This is why we invented religions, and will continue to reinvent them in ever shifting guises.
If we care about ethics, love your neighbour as yourself. Not a hypothetical neighbour on planet Zarkphobble in the year 20123894, but that smelly homeless guy in a pool of vomit outside Costco you passed this morning. And for the rest of us, let's focus on doing less harm with the technology we create.
> This is why we invented religions, and will continue to reinvent them in ever shifting guises.
Something off about this statement, though. Like out of place in some otherwise wise observations.
Man has our escapism but it isn't at all uniquely circumscribed by "religion".
Unless you define, circularly, "religion" to be man's escapism.
I'm trying to avoid some apologetics over some theology or something. For myself, I'm an apostate. Most religions people would call me an atheist, but I sense that most atheists would reject my membership application.
But what I see in essays like this is a limited human attempting the impossible task of reconstructing some meaning to existence.
The task being impossible b/c we can no more reconstruct meaning than we could reconstruct our, say, biological selves. That was the job of eons of evolutionary process.
Which is what "religion" hosts, encodes... along with crufty theology, escapism, etc.
What you seem to be pitching is this "Just be good" philosophy that for some reason I keep seeing show up on T-shirts the past couple of years. But this is not a substantive philosophy.
When I left religion in my early 20s, in retrospect I feel I threw the baby out with the bathwater not realizing that there was deep, evolved meaning and framework amid the mounds of garbage. That remained even when I thought I was rejecting all of it, the escapism, dogma, fear, etc... In other words I couldn't discard whatever was encoded in myself.
But I worry now I'm not giving my kids what I'd gotten and took for granted. Worry now that a whole generation is not getting it. That we're going to be left with "Just be good" - and otherwise these desperate, searching, doomed essays from people who realize that "Just be good" is not good enough but who have no heritage from the past that hosts pretty much all we could expect to have.
Exactly. Few things have been more pragmatic, and with more practical impact in history (both good and bad), than religion. The "fairy tale" part of it is just a distraction to the actual symbolic content.
I’m hoping someone smarter than myself can momentarily stoop down to my level and clear things up.
I wonder, must one strive to understand the mathematics of ethics to make ethical decisions? Am I able to optimise my decisions for universal wellbeing without being able to mathematically reason upon it?
Something I’m equally ill-equipped at is being present - aware of my real-time inclinations, motivations, and subsequent actions. As a result, I often make decisions which oppose ethical optimisation (or my understanding of it)
If I were to choose between the ability to mathematically reason on ethical matters, or to have the awareness to actually make (potentially misguided) ethical decisions, I’d choose the latter every time.
Is that wrong?
Once you consider every possible world and every possible outcome, happening infinitely many times, I would draw a different conclusion -- unless you have the power to effect the entire "infinite-verse", then anything you do becomes unimportant, as infinite+finite=infinite, so your contributions are lost in the wash.
So in such a dynamical system, how do you then act? The philosophically Western answer to this is in part through Logos. You treat the spoken and written word as divine, through truthful and authentic expression of the individual, as clearly as possible. The consequences of how this unfolds is outside your immediate control, as it should be.
Love has an important role here, but Love without Truth seems to be the prevailing attitude online, which is declared as a personal virtue and experienced as righteous indignation. This is to me what others perceive as the fall of the West: the prioritization of Love over Truth.
As for the infinite aspect, infinity doesn't mean uniform - there are an infinite number of odd numbers but none of them are even.
The many worlds interpretation of quantum mechanics, for instance, provides clear math on how to reason about quantities as the universe splits into more universes: we integrate after multiplying by the probability of each branch. No credible physical theory is proposing we double the amount of mass and energy in the universe whenever we measure the spin of an electron.
Calling that split universe "two universes" instead of "two half-universes" is fine for a casual explanation, but if you want to do correct math with this result, you need to start with the correct math too! Once you make that adjustment, all this talk in about "finite benefits don't matter compared to infinite benefits" can be translated as "benefits with probability zero don't matter compared to benefits with non-zero probability". Much more palatable.
There are still interesting conversations about benefits which could last for infinite time. But there, the crux is in the "could". These thought experiments always bake in assumptions at dozens of levels which say "this factor in my model will last for infinite time". When you consider in our ignorance about what the future will hold, I've yet to see a convincing example where we could truly consider a finite action as having infinite utility. I'm certainly open to it, I just haven't seen it.