ethics

> You're right that it doesn't make any difference in the pure form of the thought experiment, however I think it does make a difference when you have limited resources to build your world. It's much easier to push someone's conscious valence up than to generate another 1000 miserable beings. The main thing that makes a difference here is open/empty individualism.


I am still not sure what it is the entire point of lograthmic scaling is. The point of it being difficult to generate 1000 barely-sustainable beings applies independent of logarithmic scaling.

However, either way, there are some things I find deeply counterintuitive here.

Say I create a new human organism. I become slightly less happy as a result. Let's say the human organism itself becomes very happy. Let's say, however, the human organism doesn't do anything to the world at large to influence valence of others. Let's say I reasonably expected prior to the decision that this would happen (that I would be less happy, the created human would be very happy, and the world wouldn't change much besides that as a result).

From a sum-utilitarian POV this should be a better decision, the sum of happiness is increased, but I already find that extremely strange. I find no sense in creating a new human being (who had no interest in coming to being given they were non-existent to even have interests), at the cost of harming (even if only slightly) actual existing beings (eg. me). (I am not being against procreation in general here) even if the new being happens to be "happy".

Even in the abstract, the fact the utilitiarainism allows "just bringing new happy people" to existence to compensate for not helping existing people or even harming existing people is ridiculous to me. We don't even have to get into repugnant conclusion with high population numbers.


My concern is utilitarianism has already gone of the wrong end as soon as utility is abstracted away from persons (which is the effect of sum/avg composition). A vivid example made by Tucker that I cannot get off once made, is consider the two case:

(1) A person A has -50 hedons (2) There are 50 people with -1 hedon each.


According to vanilla sum utilitarianism case (1) and (2) is basically the same, but if we are not willing to bite big bullets out of ideological commitments that is of course far from true. It's very different of a case where a single person is suffering greatly (-50) compared to multiple people with minor annoyance (-1 hedon). The questionable "trick" of Utilitarianism here is that it has pulled appart hedons away from people to just sum it up into an abstract "overall utility of the world". The information about distribution and concentration of hedons to different people is lost. And it is not at all obvious why these information should be ignored.


It's also not obvious what "open individualism" can say to this. Even if every experience is ultimately possed by "one person", we have to at least admit that some experiences are different synchronically-and-diachronically associated to created a experiential "boundedness". And how hedons are "localized"/distributed/concentrated in different "regions" of experiences can lead to great practical differences. It's not obvious at all, that we should ignore all that for a "mere sum". These is where we have to careful with glossing concrete experiential differences through changes in linguistic frameworks (related to person individuations).


Of course, "sum utilitarianism" isn't the only utilitarianism -- there is negative utilitarianism, average utilitarianism (and I believe there is potential for utilitiarianism using better maths and formalisms) but it is non-trivial to formulate something that appears sensible to me or robust enough to hypothetical adversaries (which may be "impractical", but we don't know when and how realistic analogues can cause issues that are forewarned and clued in the hypotheticals).


There are alternative potentially better ways to design frameworks for "action ranking" than the standard utilitarianism that I come across. This may not even be a problem of utilitarianism per se but its standard mathematical formalizations stemming from (I believe) unreflected intuitions (and we may be after reflection too willing to gloss over the fundamental issues after we already become sort of committed). Still it's not easy to frame it as definitive and as the "only option"


There is a post that attempts to "solve" repugnant conclusion with logarithms (not sure if you had something similar in mind):

https://www.lesswrong.com/posts/Zgwy2QRgYBSrMWDMQ/logarithms-and-total-utilitarianism

But again I find it as no solution at all.

Precisely:

> So I have shown that there is a population M greater and better than A where everyone is worse off, how is that different from the RC? Well, the difference is that this does not happen for every population, but only for those where average well being is relatively high. Furthermore, the average individual in M is far above subsistence.

I find it already problematic that you can justify increasing population to M at the cost of existing beings (hwoever minor). Moreover, the problem canb be again worsened if we change the situation and allow resources to increase by some factor with population -- and again the numbers can be adjusted where Z is better than A counter-intuitively.


(I personally think the whole idea of calculating "utility" of the world is wrongheaded, a better framework is to "rank actions" based on concrete relative gains while taking into account costs -- and procreation needs separate careful treatment (to not allow "cheating").)


Also forget about these tradeoffs. Consider in world one, there are 5 maximmally happy people (with happiness level h each) and in world two, there are 10 maximally happy people (with happiness level h each). Even then, I have no idea why world 2 should necessarily be "better". What's so special about increased number of people? This whole premise that somehow increasing number of people is "better" flies over my head.


(Of course, going into average utilitarianism has its own problem because you can use it to justify genocides to increase averages. But again instead of this pushing me back to sums, I am suspicious of these whole paradigm).


There are also lots of other crucial issues:


For example in decision theories you have prisoner dilemma. It appears to me that a similar analogue also exists in ethics. It seems to me there is a discrepancy that arises based on how much you focus on single agents vs multi agents complexities.


For example, often we can find a situation, where if you alone do x in some context y, the consequence is better (or indifferent). But if by the same logic everyone starts to do x in context y, the consequence is a world no one may want to live in (especially if it gets known in public that chance of someone doing x if context y is high).




One example, perhaps in a circumstance y it can be better in terms of overall utility that that someone coming from a regular check up in a hospital be sacrified to save five ill patients (again I already find this kind of number games questionable to begin with). But if one person is allowed to do it for "better local consequence", why not everyone? And if everyone is allowed to do it and it is made legal, do we really want to stay in such a world (where you cannot really trust a hospital visit to not end up in your death)? Perhaps the world only seens dystopian because we are selfish and afraid of our own death but again it's start to kind of seem odd where we get into setups which no one desire for some "abstract utility function".


We can push against that with something like rule-consequentialism but if it's merely a heuristics then it is not solving the fundamental issue that such cases are allowed; if it is not just a heuristics it's start to become unclear if it is consequentialism anymore. Interestingly categorical imperative seems to vibe more with a game-theoretic consideration (when you are acting you are not just thinking locally the consequence of your own action, but you are acknowledging that your action by doing it you are implicitly treating such an act as legal -- and you should be comfortable with what follows from the universalization of your action as a law -- what happens when every agent is allowed to do the same action).


This is analogous to game-theoretic dilemmas. Where doing x, given everyone doing y (or everyone possibly doing y or x with some probability) seems advantageous, but doing x also "promotes doing x" increasing the chance of everyon doing x, and everyone doing x leads to chaos. If we then focus on acting in more game-theoretically stable manners that lead to more holistic stability then I am not sure what we get would be exactly consequentialist although not necessary far from it.


> If you go down this line of reasoning, your future self is separate from your past self, they don't share a bound experience either, it's just that your future self has a bunch of memories that it puts together into a story to relate to your past self. Most common sense ethics still tries to reduce suffering for a future self, why is this any different than claiming that you should help others?

Good. This line of reasoning is used by Santideva. But there can be multiple response.

(1) I can just allow that I don't care about my future self for some "special reason" that it is "my" future self. But I care for my future self simply because I do, and not different from reasons I care for some other people that I come across. And while I have (or grow) interest for some people by default - brute psychological forces, there still remains the open question of why I should rationally universalize my care.

(2) Another response in a similar line would be that while my future self is outside my immediate bounds, they would be still have special connections (whether we call in personal identity criteria or not) to that future self - for eg. psychological continuity and such (for example, that my future self may be more likely to carry out my will; and I can treat my future self similar to how people treat their own children).  I can simply say that I just brutely care more about people "related" to this "experience" in a certain manners (psychological continuity, history of relations) and some generic underlying metaphysical unity is not enough grounds for me to care about everyone.
.


> I mean that all consciousness is the same in the sense that two different electrons are the same, all consciousness has the same ontological properties. So if you buy that your own suffering is "real", then so is someone else's.

If you are treating different consciousnesses as only same in virtue of being the same kind (having consciousness) but still quantitatively non-identical then I am not sure what point open individualism is going to make.


Either way, I think summary (summarizing above points, and adding some extra ones):

(1) First it's not even perfectly clear to me what "consciousness valence" is and if it's what I should even care about. The question of what is "good" -- "Eudomonia" is still open. Are the Buddhist right? Is the better forms of pleasure in the lack of all pleasures/pain and mental disturbances in stilled down jnanic states like nirodha and such? I also find "pleasure" to be somewhat empty although kind of addictive. I don't have any special attraction to the notion of gain "intense pleasure" for example (even though I am not particularly happy or sad -- at least not enough to be in some logarithmic boundaries where I cannot be made much significantly happiner). Perhaps simply "raw pleasure" or "intensity" is not the right thing to look out for, perhaps we also need "meaning" or a grounding "framework", perhaps there are "higher forms of pleasure" too that are more subtle (feels less intense, but colors up experience in a way that's hard to spell and even hard to notice except in contrast to cases where there is intensities but lacking meaning). But it starts to become very nebulous and abstract for me. I am start to become unsure about what these higher pleasures are. Is being a sad Socrates really better than being a happy pig or would I be just deluding myself or is there no gap between delusion and reality when it comes to these sort of things? Or are we also supposed to look more holistically into how the agent is situated -- what relationships they have and all that? While I believe there are some good answers to these questions, I don't think all these are at all very obvious - i.e what exactly to aim for and treat as a utility for us and others. There is also the classic contest between "preference" vs "hedon". Is it really better to be forced into becoming a happy pig (if it really does increase all forms of hedons) even if I don't prefer to loose some of my hedon-indifferent features?


(2) Even if we have reached a consensus of a individual "utility", and even if we acknowledge there are real people with real "utility" values (could be real suffering), it is still a non-trivial question why a individual should care about another, if they don't brutely care for others by default for psychological reasons. It's one thing to acknowledge other people undergoing real suffering and that it is a real objective fact that there are beings who suffers and it is "bad for them", this doesn't translate to the notion of a "bad abstracted away from people", "bad not for a person but just bad". It's still a jump to make to go for a individual caring for their own utility to the utility of others. You are right that the question may need to be answered even for why one should care about one's "future selves" if they are bounded from immediate experience (is there some underlying "metaphysical entity" connecting the two selves? Should we care? Should we care if these entity is underlying every selves?). I don't think any of these is completely trivial to answer. While I think there are some decent answers, I don't think they are nearly as tautological (without making loaded linguistic moves) as you made out to be in some other comments.

(3) Even if we have reached a consensus on what kind of "utility" we ought to care about, and we have found a solid reason to care for others beyond the boundaries of immediate experience, there is still a non-trivial question about how exactly this care is to be put into practice under uncertainties and bounded resources i.e which decision policy is to be adopted. It's not obvious at all that the standard "maths" of "just sum up" or "just average up" are good answers at all. I believe we may need to go back to our boards, and start thinking from ground up decision-theoretically and game-theoretically instead of getting muddied up by current status quot, or just making ad hoc modifications to current frameworks as if they are almost correct. In fact, I find them extremely counter-intuitive from the get go, without getting into repugnant conclusion