Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Comment author: Pablo_Stafforini 10 August 2013 01:59:09PM *  1 point [-]

My general point is that whatever property you rely upon to make comparisons within pains you can also rely upon to make comparisons between pains and pleasures.

It seems to me that you are using intensity of desire to make comparisons within pains. If so, you can also use intensity of desire to make comparisons between pleasures and pains. That "there would be no reason why people need to choose an exchange rate corresponding to some measured properties" seems inadequate as a reply, since you could analogously argue that there is no reason why people should rely on those measured properties to make comparisons within pains.

However, if intensity of desire is not the property you are using to make comparisons within pains, just ignore the previous paragraph. My general point still stands: the property you are using, whichever it is, is also a property that you could use to make comparisons between pains and pleasures.

Comment author: Adriano_Mannino 16 August 2013 06:12:17PM *  1 point [-]

I'm not sure if it's possible to separate "felt intensity" from "intensity of desire". (I don't know what pain/suffering without a desire that it not exist would be.) But however that may be, your point doesn't seem to settle the population-ethical issue: If we look at hedonic desires (weighted by intensity), should we maximize [fulfilled desires - unfulfilled desires] or minimize [unfulfilled desires]? A desire can be considered a problem to be solved. If we want to solve the world's problems (which motivation seems to underly what many people are doing), does it make more sense to minimize unsolved problems or to create as many solved problems as possible? - I think clearly the former, for the non-existence of problems (and thus of solved problems) does not intrinsically pose a problem.

Why isn't it all one scale of "felt hedonic intensity"? If it was all one scale, it seems that placing the 0-point would be a purely formal and arbitrary matter. But we agree that it's not - so there seems to be something substantial going on when hedonic tone changes from "pleasurable" to "painful". We're not sliding along a scale of more/less of the same thing - at some point, the thing in question changes. Suppose I grant you that there is a way of comparing pleasure- and pain-intensities: "Here's a pain of intensity x, and there's something that's a pleasure and has the same intensity x." Now how are you going to establish that x-intensity of that other thing (pleasure) morally outweighs x-intensity of pain? Maybe it's 2x-intensity of the other thing? How's that choice not arbitrary? As Lukas said, the choice seems to be based on how much you crave the other thing (and its greater intensity), i.e. on how much of a problem its absence is to you. And this brings us back to minimizing unsolved problems, it seems.

Comment author: CarlShulman 04 August 2013 03:15:30AM 3 points [-]

I suspect that in cases people clearly recognize as altruistic, the rates are closer to Brian's than to yours.

Which cases did you have in mind?

People generally don't altruistically favor euthanasia for pets with temporarily painful but easily treatable injuries (where recovery could be followed with extended healthy life). People are not eager to campaign to stop the pain of childbirth at the expense of birth. They don't consider a single instance of torture worse than many deaths depriving people of happy lives. They favor bringing into being the lives of children that will contain some pain.

Comment author: Adriano_Mannino 11 August 2013 02:30:36AM *  4 points [-]

Sorry for the delay. - I should have been more precise. I'll provide more precision by commenting on the cases you mention:

  • The injured pet case probably involves three complications: (1) people's belief that it's an "egoistic" case for the pet (instead of it being an "altruistic" trade-off case between pet-consciousness-moments), (2) people's suffering that would result from the death/future absence of the pet, and (3) people's intuitive leaning towards an (ideal) preference view (the pet "wants" to survive, right? - that would be compatible with a negative(-leaning) view in population ethics according to which what (mainly) matters is the avoidance of unfulfilled preferences).

  • It's clear that evolutionary beings will have a massive bias against the idea of childbirth being morally problematic (no matter its merit). Also, people would themselves be suffering/have thwarted preferences from childlessness.

  • People consider death an "egoistic" case - a harm to the being that died. I think that's confused. Death is the non-birth of other people/people-moments.

  • People usually don't "favor" it in the sense of considering it morally important/urgent/required. They tend to think it's fine if it's an "OK deal" for the being that comes into existence (here again: "ego"-case confounder). By contrast, they think it's morally important/urgent not to bring miserable children into existence. And again, we should abstract from childbirth and genealogical continuation (where massive biases are to be expected). So let's take the more abstract case of Omelas: Would people be willing to create miserable lives in conjunction with a sufficiently great number of intensely pleasurable lives from scratch, e.g. in some remote part of the universe (or in a simulation)? Many would not. With the above confounders ("egoism" and "personal identity", circumstantial suffering, and a preference-based intuitive axiology) removed and the population-ethical question laid bare, many people would not side with you. One might object: Maybe they agree that Omelas is much more intrinsically valuable than non-existence, but they accept deontological side-constraints against actively causing miserable lives, which is why they can't create it. But in that case they shouldn't prevent it if it arose naturally. Here again, though, my experience is that many people would in fact want to prevent abstract Omelas scenarios. Or they're at least uncertain about whether letting even Omelas (!) happen is OK - which implies a negative-leaning population axiology.

Comment author: Pablo_Stafforini 04 August 2013 02:44:55PM 0 points [-]

That's incorrect. David Pearce claims that pains below a certain intensity can't be outweighed by any amount of pleasure. Both Brian and Dave agree that Dave is a (threshold) negative utilitarian whereas Brian is a negative-leaning utilitarian.

Comment author: Adriano_Mannino 10 August 2013 03:27:13PM *  2 points [-]

Not so sure. Dave believes that pains have an "ought-not-to-be-in-the-world-ness" property that pleasures lack. And in the discussions I have seen, he indeed was not prepared to accept that small pains can be outweighed by huge quantities of pleasure. Brian was oscillating between NLU and NU. He recently told me he found the claim convincing that such states as flow, orgasm, meditative tranquility, perfectly subjectively fine muzak, and the absence of consciousness were all equally good.

Comment author: CarlShulman 03 August 2013 07:36:31AM 3 points [-]

Lots of dystopian future scenarios are possible. Here are some of them. But even if you do believe that non-existence poses a moral problem and creates an urgency to act, it's not clear yet that the value of the future is net positive.

Even Brian Tomasik, the author of that page, says that if one trades off pain and pleasure at ordinary rates the expected happiness of the future exceeds the expected suffering, by a factor of between 2 and 50.

Comment author: Adriano_Mannino 04 August 2013 02:07:43AM *  1 point [-]

Regarding "people's ordinary exchange rates", I suspect that in cases people clearly recognize as altruistic, the rates are closer to Brian's than to yours. In cases they (IMO confusedly) think of as "egoistic", the rates may be closer to yours. - This provides an argument that people should end up with Brian upon knocking out confusion.

Comment author: Pablo_Stafforini 03 August 2013 11:44:17AM *  2 points [-]

It is worth pointing out that by 'insane', Brian just means 'an exchange rate that is very different from the one I happen to endorse.' :-) He admits that there is no reason to favor his own exchange rate over other people's. (By contrast, some of these other people would argue that there are reasons to favor their exchange rates over Brian's.)

Comment author: Adriano_Mannino 04 August 2013 02:00:33AM *  1 point [-]

Also, still others (such as David Pearce) would argue that there are reasons to favor Brian's exchange rate. :)

Comment author: Nick_Beckstead 03 August 2013 08:34:46AM 1 point [-]

Do you not argue for that at all? I thought you were going in the direction of establishing an axiological and deontic parallelism between the "wretched child" and the "happy child".

I do some of that in chapter 4. I don't engage with speculative arguments that the future will be bad (e.g. the dystopian scenarios that negative utilitarians like to discuss) or make my case by appealing to positive trends of the sort discussed by Pinker in Better Angels. Carl Shulman and I are putting together some thoughts on some of these issues at the moment.

The quoted passage ("all potential value is found in [the existence of] the well-being of the astronomical numbers of people who could populate the far future") strongly suggests a classical total population ethics, which is rejected by negative utilitarianism and person-affecting views. And the "therefore" suggests that the crucial issue here is time preference, which is a popular and incorrect perception.

Maybe so. I think the key is how you interpret the word "value." If you interpret as "only positive value" then negative utilitarians disagree but only because they think there isn't any possible positive value. If you interpret it as "positive or negative value" I think they should agree for pretty straightforward reasons.

Comment author: Adriano_Mannino 04 August 2013 01:52:55AM 6 points [-]

One important reason they like to discuss them is the fact that many people just assume, without adequate consideration and argument, that the future will be hugely net positive. Which comes at no surprise, given the existence of relevant biases.

Whether negative utilitarians believe that "there isn't any possible positive value" is semantics, I think. The framing you suggest is probably a semantic (and thus bad) trigger of the absurdity heuristic. With equal substantial justification and more semantic charity, one could say that negative utilitarians believe that the absence of suffering/unfulfilled preferences or suffering-free world-states have positive value (and one may add either that they believe that the existence of suffering/unfulfilled preferences has negative value or that they believe there isn't any possible negative value).

Comment author: Wei_Dai 19 July 2013 06:30:58AM 6 points [-]

But one may disagree on "Omelas and Space Colonization", i.e. on how many lives worth living are needed to "outweigh" or "compensate for" miserable ones (which our future existence will inevitably also produce, probably in astronomical numbers, assuming astronomical population expansion).

A superintelligent Singleton (e.g., FAI) can guarantee a minimum standard of living for everyone who will ever be born or created, so I don't understand why you think astronomical population expansion inevitably produces miserable lives.

Also, I note that space colonization can produce an astronomical number of QALYs even assuming no population growth, by letting currently existing people continue to live after all the negentropy in our solar system has been exhausted.

Comment author: Adriano_Mannino 03 August 2013 06:55:49AM *  8 points [-]

Yes, it can. But a Singleton is not guaranteed; and conditional on the future existence of a Singleton, friendliness is not guaranteed. What I meant was that astronomical population expansion clearly produces an astronomical number of most miserable, tortured lives in expectation.

Lots of dystopian future scenarios are possible. Here are some of them.

How many happy people for one miserable existence? - I take the zero option very seriously because I don't think that (anticipated) non-existence poses any moral problem or generates any moral urgency to act, while (anticipated) miserable existence clearly does. I don't think it would have been any intrinsic problem whatsoever had I never been born; but it clearly would have been a problem had I been born into miserable circumstances.

But even if you do believe that non-existence poses a moral problem and creates an urgency to act, it's not clear yet that the value of the future is net positive. If the number of happy people you require for one miserable existence is sufficiently great and/or if dystopian scenarios are sufficiently likely, the future will be negative in expectation. Beware optimism bias, illusion of control, etc.

Comment author: lukeprog 19 July 2013 06:18:42AM 1 point [-]

Okay, so we're talking about two points: (1) whether current people have more value than future people, and (2) whether it would be super-good to create gazillions of super-good lives.

My sentence mentions both of those, in sequence: "Many EAs value future people roughly as much as currently-living people [1], and think that nearly all potential value is found in the well-being of the astronomical numbers of people who could populate the far future [2]..."

And you are suggesting... what? That I switch the order in which they appear, so that [2] appears before [1], and is thus emphasized? Or that I use your phrase "morally urgent to" instead of "nearly all potential value is found in..."? Or something else?

Comment author: Adriano_Mannino 03 August 2013 06:23:33AM *  5 points [-]

Sorry for the delay!

I forgot to clarify the rough argument for why (1) "value future people equally" is much less important or crucial than (2) "fill the universe with people" here.

If you accept (2), you're almost guaranteed to be on board with where Bostrom and Beckstead are roughly going (even if you valued present people more!). It's hardly possible to then block their argument on normative grounds, and criticism would have to be empirical, e.g. based on the claim that dystopian futures may be likelier than commonly assumed, which would decrease the value of x-risk reduction.

By contrast, if you accept (1), it's still very much an open question whether you'll be on board.

Also, intrinsic time preference is really not an issue among EAs. The idea that spatial and temporal distance are irrelevant when it comes to helping others is a pretty core element of the EA concept. What is an issue, though, is the question of what helping others actually means (or should mean). Who are the relevant others? Persons? Person-moments? Preferences? And how are they relevant? Should we ensure the non-existence of suffering? Or promote ecstasy too? Prevent the existence of unfulfilled preferences? Or create fulfilled ones too? Can you help someone by bringing them into existence? Or only by preventing their miserable existence/unfulfilled preferences? These issues are more controversial than the question of time preference. Unfortunately, they're of astronomical significance.

I don't really know if I'm suggesting any further specific change to the wording - sorry about that. It's tricky... If you're speaking to non-EAs, it's important to emphasize the rejection of time preference. But there shouldn't be a "therefore", which (in my perception) is still implicitly there. And if you're speaking to people who already reject time preference, it's even more important to make it clear that this rejection doesn't imply "fill the universe with people". One solution could be to simply drop the reference to the (IMO non-decisive) rejection of time preference and go for something like: "Many EAs consider the creation of (happy) people valuable and morally urgent, and therefore think that nearly all potential value..."

Beckstead might object that the rejection of heavy time preference is important to his general conclusion (the overwhelming importance of shaping the far future). But if we're talking that level of generality, then the reference to x-risk reduction should probably go or be qualified. For sufficiently negative-leaning EAs (such as Brian Tomasik) believe that x-risk reduction is net negative.

Perhaps the best solution would be to expand the section and start by mentioning how the (EA-uncontroversial) rejection of time preference is relevant to the overwhelming importance of shaping the far future. Once we've established that the far future likely dominates, the question arises how we should morally affect the far future. Depending on this question, very different conclusions can result e.g. with regard to the importance and even the sign of x-risk reduction.

Comment author: Nick_Beckstead 20 July 2013 06:04:24PM 2 points [-]

It's not really clear to me that negative utilitarians and people with person-affecting views need to disagree with the quoted passage as stated. These views focus primarily on the suffering aspect of well-being, and nearly all of the possible suffering is found in the astronomical numbers of people who could populate the far future.

To elaborate, in my dissertation, I assume--like most people would--that a future where humans have great influence would be a good thing. But I don't argue for that and some people might disagree. If that's the only thing you disagree with me about, it seems you actually still end up accepting my conclusion that what matters most is making humanity's long-term future development go as well as possible. It's just that you end up focusing on different aspects of making the long-term future development go as well as possible.

Comment author: Adriano_Mannino 03 August 2013 04:22:52AM *  6 points [-]

Hi Nick, thanks! I do indeed fully agree with your general conclusion that what matters most is making our long-term development go as well as possible. (I had something more specific in mind when speaking of "Bostrom's and Beckstead's conclusions" here, sorry about the confusion.) In fact, I consider your general conclusion very obvious. :) (What's difficult is the empirical question of how to best affect the far future.) The obviousness of your conclusion doesn't imply that your dissertation wasn't super-important, of course - most people seem to disagree with the conclusion. Unfortunately and sadly, though, the utility of talking about (affecting) the far future is a tricky issue too, given fundamental disagreements in population ethics.

I don't know that the "like most people would" parenthesis is true. (A "good thing" maybe, but a morally urgent thing to bring about, if the counterfactual isn't existence with less well-being, but non-existence?) I'd like to see some solid empirical data here. I think some people are in the process of collecting it.

Do you not argue for that at all? I thought you were going in the direction of establishing an axiological and deontic parallelism between the "wretched child" and the "happy child".

The quoted passage ("all potential value is found in [the existence of] the well-being of the astronomical numbers of people who could populate the far future") strongly suggests a classical total population ethics, which is rejected by negative utilitarianism and person-affecting views. And the "therefore" suggests that the crucial issue here is time preference, which is a popular and incorrect perception.

Comment author: SaidAchmiz 28 July 2013 10:18:25PM 0 points [-]

Sure. However, you raise what is in principle a very solid objection, and so I would like to address it.

Let's say that I would, all else being equal, prefer that a dog not be tortured. Perhaps I am even willing to take certain actions to prevent a dog from being tortured. Perhaps I also think that two dogs being tortured is worse than one dog being tortured, etc.

However, I am willing to let that dog, or a million dogs, or any number of dogs, be tortured to save my grandmother from the same fate.

What are we to make of this?

In that case, some component of our utilitarianism might have to be re-examined. Perhaps dogs have a nonzero value, and a lot of dogs have more value than only a few dogs, but no quantity of dogs adds up to one grandmother; but on the other hand, some things are worth more than one grandmother (two grandmothers? all of humanity?).

Real numbers do not behave this way. Perhaps they are not a sufficient number system for utilitarian calculations.

(Of course, it's possible to suppose that we could, if we chose, construct various hypotheticals (perhaps involving some complex series of bets) which would tease out some inconsistency in that set of valuations. That may be the case here, but nothing obvious jumps out at me.)

Comment author: Adriano_Mannino 28 July 2013 10:43:17PM 8 points [-]

What about a random human instead of your grandmother? What if the human's/your grandmother's cognitive capacities were lower than the dog's or the chimp's? – What would a good altruist do?

How do you block the "chain of comparables"?

View more: Next