I know I went back and forth on this issue in the past but I think I've finally come to a conclusion that I think makes clear sense.

The only point I disagree with (hedonistic) utilitarians on is the idea that the happiness/suffering of separate beings can be objectively aggregated. Happiness can be measured in terms of intensity and duration but it can't be objectively quantified because it doesn't exist in objective reality, it's an abstraction. Happiness is a subjective experience, it's meaningless to talk about aggregating an experience because an experience is subjectively felt and there is no shared consciousness between two or more beings. Some pro-aggregation HUs will concede, even if they insist it will never be a practical concern, that it might be theoretically justifiable to cause one person 1000 points of pain to prevent 1000 people from experiencing 2 points of pain each but if the morally right decision to make in any scenario is the one that maximizes the greatest happiness/minimizes the most suffering (I think it is), this makes no sense because there is no subjectively felt 2000 points of stress that's being prevented that would outweigh the man's subjectively felt 1000 points of pain. You haven't decreased the greatest suffering, by causing this one person 1000 points of pain to prevent 1000 other people from experiencing a much less amount of pain, you've decreased the number of people who suffer. The goal of ethics shouldn't be to increase/decrease the greatest number of happy or distressed people but to increase/decreased the greatest subjectively felt happiness/suffering.

The argument comparing the 2 points of stress the 1000 separate people would feel to feeling 2 points of pain a day for 1000 day is what prevented me from coming to this conclusion earlier but I no longer think it holds any weight because it still depends on the idea that there is some kind of supermind that experiences the collective 2000 points of pain intermittently, if not at once. A pinprick every day for 1000 days, with the expectation that you'll xperience it again later or putting it into context with how many times you've experienced it in the past is different than a pin prick you experience once alone. If you wipe out my memory so I longer remember having experienced the pin prick the day before or have any expectation of experiencing it again in future, my argument would still stand, that there's no subjectively felt 2000 points of pain that would make it any worse than experience 1000 points of pain at once.

Intuition aside, it's difficult to take the idea that a universe with 10 billion moderately happy beings could be better than a universe with 1 billion extremely happy beings.

I'm open to opposing arguments but I think utilitarianism would be a lot easier to swallow for many people once you remove the idea of aggregation.

So, in order to avoid aggregation, you're simply using the happiness of one person at one moment. You seem to have chosen the one with the most suffering. Why? Why not the happiest, for example?

I notice that if you accept this idea, the best thing to do would be to end the world, so you don't end up with someone worse than the worst who already exists.

so kinda a negative utilitarian?:) i think you have your points wrong because if a billion people being pricked by a needle every day is -1billion id say that's equal to someone feeling a bit lonely for a day

First, here are some relevant links:
Peter Singer: Should we Ration Healthcare?
Katja Grace: Is it Repugnant?
Felicifia: Average vs Total [& the repugnant conclusion]
Felicifia: Extreme and Minor Suffering [& the human ability to think quantitatively]

Secondly, I'll instantly grant your point that aggregation is unintuitive. This doesn't just bother utilitarians. It's a problem for all economists. Aggregation is unintuitive because human beings are awful at thinking quantitatively. We look at the happiness of two different human beings, and we can ponder the positive and negative characteristics of each person's life situation for hours, without getting anywhere.

Most people don't like maths. But sane people recognise that maths is true despite that. Likewise, aggregation makes people nervous. But, you must at least consider that this might be a problem with people, not a problem with aggregation.

Thirdly, I have to contest one of your main ideas. You've said that happiness is abstract. To me, happiness is anything but abstract. It's the realest thing in the world, because it's part of my consciousness, even though noone else can see it.

You feel uncomfortable aggregating the happiness of different people. But, I imagine, you'd be comfortable aggregating the happiness of one person across time. So, as DanielLC said in another thread, you'd be content to say that if a person saves money to throw a birthday party, that party can truly be successful or unsuccessful, and that there is something that it means for a saving to be worthwhile. Like most people, you'll say that's because of something called identity that is bounded by space, but not by time, right? No, now you're abstracting. To me at least, identity is an idea that's far harder to pin down than happiness. Is that an approach that you can relate to at all?

Secondly, I'll instantly grant your point that aggregation is unintuitive. This doesn't just bother utilitarians. It's a problem for all economists. Aggregation is unintuitive because human beings are awful at thinking quantitatively. We look at the happiness of two different human beings, and we can ponder the positive and negative characteristics of each person's life situation for hours, without getting anywhere.

Most people don't like maths. But sane people recognise that maths is true despite that. Likewise, aggregation makes people nervous. But, you must at least consider that this might be a problem with people, not a problem with aggregation.

It's not the counter-intuitiveness of aggregation or the fact that it makes me 'uncomfortable' that's lead me to this conclusion, I believe it's irrational. Simply put, happiness is a *subjective* experience, you can't objectively aggregate a *subjective* experience because subjective experience doesn't exist in the objective world the way that water or rocks or any other empirically observable substance does. If there are 5 rocks in a container and you add another rock, you now have 6 rocks, rocks can be aggregated because they are empirically observable objects. If there are 5 happy people in a room and you add another, you now have 6 happy people but you haven't increased the actual happiness in the room (happiness only exists in individuals, not in 'rooms', and each individual 'mind' is a separate mental universe), you've increased the number of happy people. A human, as an organism, is empirically observable and humans can be aggregated because of this, an experience cannot be empirically observed so it cannot be aggregated. Maximizing happiness means maximizing the intensity or duration of any individual's happiness, since individuals, not groups or the universe, feel happiness.

If given the opprtunity to give one person stomach cancer or to punch someone else in the nose, I would choose the latter because I can imagine the pain of being punched in the nose and compare it to the imagined pain of having stomach cancer. I cannot imagine the 'collective' pleasure of 1000 people because it doesn't exist, you can't add their pleasure in the way that you can add 2 litres of water to another 2 litres. If you could add their separate pleasure, they wouldn't be separate 'minds'. I can say, from the point of view of the would be stomach cancer victim, that giving him stomach cancer would be worse than giving the other fellow a punch in the nose. I cannot say, from the point of view of those 1000 people, that giving each of them a point of stress would be worse than giving one individual 100 felt points of stress because those 1000 people don't have a collective point of view and neither does the 'universe' or the 'world'.

Thirdly, I have to contest one of your main ideas. You've said that happiness is abstract. To me, happiness is anything but abstract. It's the realest thing in the world, because it's part of my consciousness, even though noone else can see it.

Happiness is real, by 'abstract', I just mean that it doesn't refer to anything concrete or objective. In the same way that 'freedom' refers to an abstract state of affairs and not something we can drink, eat, listen to etc.

You feel uncomfortable aggregating the happiness of different people. But, I imagine, you'd be comfortable aggregating the happiness of one person across time.

Not necessarily.

So, as DanielLC said in another thread, you'd be content to say that if a person saves money to throw a birthday party, that party can truly be successful or unsuccessful, and that there is something that it means for a saving to be worthwhile.

If the pleasure the party causes would outweigh the distress/pleasure that saving causes or prevents at any given moment, then it is worthwhile.

Like most people, you'll say that's because of something called identity that is bounded by space, but not by time, right? No, now you're abstracting. To me at least, identity is an idea that's far harder to pin down than happiness. Is that an approach that you can relate to at all?

I'm not sure what you mean. If I were to experience 1 point of stress every day for 1000 days but my memory was wiped out everytime it happened, there would be no subjectively felt 1000 points of stress (that would be true even if my memory wasn't wiped except the expection of experiencing it again tomorrow along with the memory of having experienced it yesterday would add additional stress). It would be just as wrong to cause me a felt 100 points of stress in order to prevent a not-collectively felt 1 point of stress for 1000 days.

As for identity, consciousness is a moment to moment experience. I think 'we' stop existing during non-REM sleep (assuming we're unconscious during NREM sleep). Our memories are what give us the illusion of a continued existence. So I think I'm willing to say that, besides the pain/pleasure we get from remembering past experiences or the expectation of pleasure/pain we will feel in future, you can't aggregate the total pleasure/pain of an individual's life either.

DanielC,

I'm not sure I understand your first paragraph

so kinda a negative utilitarian?:) i think you have your points wrong because if a billion people being pricked by a needle every day is -1billion id say that's equal to someone feeling a bit lonely for a day

I believe that there is a *practical* assymetry between pleasure and pain, since animals feel pain more easily than we feel pleasure, but there is a theoretical amount of pleasure that is equivalent to the worst suffering in the world.

I hope this goes through, my computer will break down any minute now.

Ubuntu wrote:If there are 5 rocks in a container and you add another rock, you now have 6 rocks, rocks can be aggregated because they are empirically observable objects.

This isn't immediately an argument for aggregation, but this remark is basically false. A 'rock' is a collection of atoms that we treat as being similar enough to certain other collections of atoms (though possibly excluding any atoms of the same type) to bracket them into the same category. We do this because it can be helpful to us to rely on analogies. I want to bash my prey with something that behaves much like the thing I bashed my last prey with. But there's no essential sameness between two rocks, or any other two objects bigger than protons.

So you can't actually aggregate *anything* that's big enough for you to see. Conversely, it's impossible for us to get by without treating certain things as though they were similar enough to add to. Happiness seems to me to be one of those things.

For another negative thought experiment, can you give any reason to say why happiness even within one person should be essentially cumulative, given modern neuroscience? If not, how can you justify making any decisions based on the idea that you want thing A more than thing B?

Happiness doesn't objectively exist in the sense that we can't just build a hedonometer and measure it, but that doesn't really matter. Suppose there happens to be some particle that isn't affected by any of the four forces. It would be completely impossible to detect, but so what? You could still have exactly five of them.

I am not a hedonistic utilitarian, but rather a more general kind of welfarist utilitarian - I think happiness, preference satisfaction, Sen's capabilities, etc. are all important elements of welfare. Therefore any prediction about the likely impact of an action on aggregate welfare is not only an empirical matter - it also depends on how much weight you put on different elements of welfare.

But I think this is even the case for a hedonistic utilitarian. We can't objectively say what the value of the physical pleasure of eating is compared to intellectual pleasure compared to a feeling of satisfaction with one's life. A naive view of hedonistic utilitarianism might say that once we decide maximizing aggregate happiness is the goal, how we get there is an empirical matter. I think this view is incorrect. Rather, I think that what weight we put on different forms of happiness is a matter of values as well - there is no objective empirical truth of the matter about what the value of, say, mystical experience is compared to romantic love.

However, both for hedonistic and a more general welfarist utilitarianism, we do need to put weight on different kinds and amounts of happiness in different people. We need to say if a loss of this kind of welfare for these people is outweighed by a gain in this other kind of welfare for these other people. This is partly an empirical matter, and partly a value judgement (through the latter is heavily informed by empirical information). There are many cases in which it is intuitively clear what kinds and amounts of welfare we should value over others, and many where it is more ambiguous. This means that utilitarianism more ambigous than we would like it to be, but in the end it is the best we have.

Finally, I do oppose "total" utilitarianism, which says we should maximize aggregate welfare period. Rather, I am an advocate of prior-existant utilitarianism. This means that when making a decision, my goal is to maximize the aggregate welfare of all people who will exist independent of what decision I make (prior-existant people). So bringing about the existence of more people is not good or bad independent of its effect on people who already exist, or will exist regardless of what my decision is. This view avoids Parfit's repungent conclusion (so a universe with 10 billion moderately happy beings is not necessarily better than a universe with 1 billion extremely happy beings), but it keeps the idea of aggregation of welfare.

This isn't immediately an argument for aggregation, but this remark is basically false. A 'rock' is a collection of atoms that we treat as being similar enough to certain other collections of atoms (though possibly excluding any atoms of the same type) to bracket them into the same category. We do this because it can be helpful to us to rely on analogies. I want to bash my prey with something that behaves much like the thing I bashed my last prey with. But there's no essential sameness between two rocks, or any other two objects bigger than protons.

I thought someone might mention this but we can still empirically observe rocks. Even a quantum particle is still an empirically observable 'thing', an experience is not. Experience only exists for the experiencer.

So you can't actually aggregate *anything* that's big enough for you to see. Conversely, it's impossible for us to get by without treating certain things as though they were similar enough to add to. Happiness seems to me to be one of those things.

I'm just not convinced by this.

For another negative thought experiment, can you give any reason to say why happiness even within one person should be essentially cumulative, given modern neuroscience? If not, how can you justify making any decisions based on the idea that you want thing A more than thing B?

If I understand what you're trying to say, I don't think I can. 5 very stressful days (let's say the stress is due to something specific like working or dealing with some crises) would be no worse than 1 stressful day if it weren't for the memory of having previously had a stressful day or the expectation of another stressful day in future.

I can imagine what it would be like to feel 60 points of pleasure for 1 hour but it's so meaningless to compare my feeling 60 points of pleasure for 1 hour to 60 people experiencing 1 point of pleasure for 1 minute *or* to my experiencing 1 point of pleasure for 1 minute intermittently and having my memory erased after every 'session'. There is no subjectively felt 60 points of pleasure, there's a felt 1 point of pleasure that occurs many different times. Who wants to eat single skittle every day for 365 days rather than 100 skittles at once?

From whose point of view is 100 separate people experiencing 1 point of pleasure as good as one person experiencing 50 points of pleasure (in a specific moment, I concede that aggregating the pleasure/pain of a single individual over days/months/years etc., when you don't take into consideration memories or future expectations, makes little sense as well)? If the 100 people don't have a 'collective' point of view and the universe as a whole doesn't have a 'point of view' either, and none of those individuals benefit from 1 point of pleasure as much as the person who experiences 50 points of pleasure, where is the beneficiary?

1 million chimpanzees will never be able to understand what Stephen Hawkins, as one individual does, because 1 million chimpanzees don't have a collective IQ.

It would be completely impossible to detect, but so what? You could still have exactly five of them.

But happiness isn't a thing, it's an experience. Neurons are things, neurotransmitters are also 'things', the brain's subjective perception of chemical information is not a 'thing'.

It occurs to me that the same basic argument can be used against expected utility. If there's a 50% chance of getting one point of pleasure, and a 50% chance of getting three points of pleasure, nobody is getting two points of pleasure. Do you treat it as one, two, or three points?

So what is the upshot of this for you Ubuntu? Would you agree with what I said above, that while happiness can't be aggregated in an "objective" sense, that utilitarians can attach different weights to different experiences of happiness/weflare based on our (subjective) values, and use these weights to (very roughly) determine the impact of a decision on aggregate value-weighted happiness/welfare?

If not, and if you do agree with utilitarians in most other points, how would you go about making moral decisions? It seems that in order to make a decision based on consequentialism you almost always need some way to weigh the different consequences of your actions. Unless your action causes all good consequences and no bad consequences (to the degree that makes sense), you are going to have to decide if the good consequences outweigh the bad, or vice versa.

It occurs to me that the same basic argument can be used against expected utility. If there's a 50% chance of getting one point of pleasure, and a 50% chance of getting three points of pleasure, nobody is getting two points of pleasure. Do you treat it as one, two, or three points?
-------------------------------------------------------------------------------------------------------------------------------------------

My dearest Daniel,

I don't see how it can. Either X will cause 5 points of pleasure or it won't. There isn't actually a "chance" that it will or won't, it either will or it won't but we have to work with the knowledge that we have so we estimate that it probably will or won't even though it actually will or won't. I'm not sure what you're saying but expected pleasure isn't pleasure.

"So what is the upshot of this for you Ubuntu? Would you agree with what I said above, that while happiness can't be aggregated in an "objective" sense, that utilitarians can attach different weights to different experiences of happiness/weflare based on our (subjective) values, and use these weights to (very roughly) determine the impact of a decision on aggregate value-weighted happiness/welfare?

If not, and if you do agree with utilitarians in most other points, how would you go about making moral decisions? It seems that in order to make a decision based on consequentialism you almost always need some way to weigh the different consequences of your actions. Unless your action causes all good consequences and no bad consequences (to the degree that makes sense), you are going to have to decide if the good consequences outweigh the bad, or vice versa. "

Beloved Brent,

I'm 'slow' so I have no clue what you're trying to say but I reject pluralistic theories of value. I don't see what shared properties happiness/pleasure has in common with other 'forms' of 'welfare' to make them all inherently valuable. I believe that ethics should be patient centered and that actual consequences (how a decision emotionally affects everyone impacted by it) determine the rightness or wrongness of a decision, not our subjective values (again, I'm not sure what you're trying to say).

When it comes to determining which actions will probably increase/decrease happiness or suffering, I would use cognitive empathy, basic reasoning, imagination and intuition. I'm willing to cause one person 10 points of stress ('points' in terms of intensity/duration) in order to prevent 20 points of stress in someone else because the felt 10 points is still not as bad as the felt 20 points, it is the lesser of two evils. I'm not willing to cause one person 10 points of stress to prevent 20 people from experiencing 1 point of stress each because there is no collectively felt 20 points of stress that is worse than the actually felt 10 points of stress. This is where empathy comes in, the idea of aggregating pleasure/pain is counter intuitive because it conflicts with our innate sense of empathy. I avoided mentioning this initially because whether or not the idea of aggregating pleasure/pain is compatible with empathy has nothing to do with whether or not it's logical but, for the record, it absolutely is not compatible with empathy because there's no super-mind whom we can empathize with that feels the collective pleasure/pain of 2 or more people, in the same way that 5 mentally retarded people with a collective IQ of 200 can't understand, 'collectively', what an individual with an IQ of 150 can. The starting point of empathy is being able to imagine the internal, emotional experience of other beings, if an action causes 20 people to experience 1 point of pleasure each, I can't imagine their collective 20 points of pleasure because there is no felt 20 points of pleasure to imagine. You can compare it to experiencing 1 point of pleasure for 20 days but this isn't applicable because the the memory of having experienced a point of pleasure each day before or the expectation of experiencing another point of pleasure the next day and the next etc. would be a source of pleasure beyond the actual pleasure being caused, none of the individuals experiencing 1 point of pleasure alone have any connection to the pleasure that the others feel.

Riveting Ryan said earlier that aggregation makes me uncomfortable because, like most people, I'm not a very mathematical person (it's true, I'm not but that's not why I no longer accept it). I can understand why 'objective' thinkers/left brain personality types (extroverts, sensors, thinkers, judgers, I'm also probably a judger ) would be drawn toward the idea of aggregation because they're used to dealing with the external, visible, objective world but this kind of mathematical thinking doesn't apply to happiness/suffering because, like I said, it doesn't exist in the objective world, it's an internal perception.

Ubuntu wrote: . . . A pinprick every day for 1000 days, with the expectation that you'll xperience it again later or putting it into context with how many times you've experienced it in the past is different than a pin prick you experience once alone. If you wipe out my memory so I longer remember having experienced the pin prick the day before or have any expectation of experiencing it again in future, my argument would still stand, that there's no subjectively felt 2000 points of pain that would make it any worse than experience 1000 points of pain at once. . .

Ubuntu, I think you're on to something pretty good. Our gut instinct (well-tutored by experience) is that intense suffering counts somewhat more. And we need not make the move which sometimes follows that this means (infinite) negative utilitarianism, which means destroy the world. Not necessarily at all. Why can’t we just say that intense suffering counts somewhat more? I think we can. And we can call this a moderate version of negative utilitarianism. Or, this might just be plain old negative utilitarianism and more of what Karl Popper originally had in mind.

You may have heard the rejoinder to a lot of traditional economics that it’s better to be approximately correct than exactly mistaken. You may have even heard the joke where a guy is floating in a balloon and he’s lost. There’s a break in the clouds and he floats down toward the Earth and a person is there. He cups his hands and yells to the person, “Hi. Can you tell me where I’m at?” The person yells back, “You are in a balloon.” Now, depending on the joke, the person on the ground is either an economist or a Microsoft employee, in that the answer is completely accurate but totally uninformative.

So, why can't we just say what would typically count as 2000 level suffering has disutility of 3,500? Now, we're not at all sure that's exactly correct. But as a disutility which is almost double, but not quite, I’d say it’s in the ballpark.

------------------------------------------------------------------

And then there’s the social justice question: Is it a good idea to impinge on one person for the benefit of the group? Generally not, and we know this pretty well from human history, and that idea might be kind of present in backs of our minds in response to various proposals of aggregation. Impinging on one person is typically a low-percentage move. And obviously, we’d prefer high-percentage moves. I’d even say, we want a series of medium-risk, high percentage moves, with monitoring feedback as we go along. That way, we get more of a dynamic process going.

Hi Ubuntu, thanks for your reply

Ubuntu wrote:I'm 'slow' so I have no clue what you're trying to say but I reject pluralistic theories of value. I don't see what shared properties happiness/pleasure has in common with other 'forms' of 'welfare' to make them all inherently valuable. I believe that ethics should be patient centered and that actual consequences (how a decision emotionally affects everyone impacted by it) determine the rightness or wrongness of a decision, not our subjective values (again, I'm not sure what you're trying to say).

I'm sorry I see I wasn't very clear. Hopefully I can be clearer now:

I believe that ethics should be focused on actual consequences for the welfare of people or animals (not sure what "patient centered" means), but I also believe that ethical values are subjective. This is why I don't see a problem with being a pluralist about what counts as welfare - I value happiness, but I also value preference satisfaction and capabilities as well.

I'm willing to cause one person 10 points of stress ('points' in terms of intensity/duration) in order to prevent 20 points of stress in someone else because the felt 10 points is still not as bad as the felt 20 points, it is the lesser of two evils. I'm not willing to cause one person 10 points of stress to prevent 20 people from experiencing 1 point of stress each because there is no collectively felt 20 points of stress that is worse than the actually felt 10 points of stress.

I agree that there is no collectively felt 20 points of stress, but that doesn't mean it isn't worse than an actually felt 10 points of stress. I also don't think there is an objective way to compare the pain of two different people, especially if the pains are of different types. How can you compare one person's deep existential angst with another person's feeling of betrayal by their best friend? How would you compare these with a third person's broken bone?

Any moral decision comes down to value judgements, but just saying that we are utilitarians - and so value happiness - is not enough. Every time we compare consequences for different people's happiness (or wellbeing), we are making value judgements about how much weight we put on different experiences of happiness or unhappiness.

Finally, here are a couple of questions for your view: (1) If you had the choice of (a) saving the lives of 5 people, or (b) saving the lives of 2 people, would you have any reason to choice a over b? If so, how could you justify this without aggregation? (2) You said that it is impossible to aggregate pains in a single person across time. Does that mean that the duration of a pain doesn't matter, just its intensity?

Brent wrote: . . . This is why I don't see a problem with being a pluralist about what counts as welfare - I value happiness, but I also value preference satisfaction and capabilities as well. . .

And whether something is a direct good or "merely" indirect, I don't know if it makes that much difference as far as building teams and moving forward with big improvements, which I think we should be about. And I might also add authenticity and friendship as part of our pluralistic goods. Sure, they're indirect goods, but as far as the world we actually live in, they tend to have big payoff (not always, of course!).

So, I suppose the shorthand I favor would be: moving beyond first principles to social skills. But I think there's a lot more to it than just that.

For example, as I understand it, almost any school reform "works" as long as the teachers believe in it, perhaps because they then bring fresh energy. So, what an administrator does not want to do is to try and shoove a program down teachers' throats. Instead, lightly skim, find a program teachers believe in, roll with that. Get feeback, roll with a second set of reforms, more feedback, a third set. So, we can be at the 5th generation, we can be ping-ponging back and forth between theory and practice that much, in the same time that we can complete one so-called perfect study.

DanielLC wrote:Happiness doesn't objectively exist in the sense that we can't just build a hedonometer and measure it, but that doesn't really matter.

Ah, but we can! For pain, at least.

And now I've actually found a more serious approach to the hedonometer idea, using statistical analysis of affective tone in many blogs:

http://www.physorg.com/news167661927.html

If I understand the OP's argument correctly, it would follow that most if not all of the work done in the fields of psychology and psychiatry is invalid, as it relies on aggregating subjective mental states. For example, you test if a pill or form of psychotherapy for depression works by randomly assigning people to treatment and control groups, measuring and aggregating the depression scores after treatment of members of each group, and comparing those aggregated scores. This anti-aggregation argument is one that Richard Ryder, ironically a psychologist himself, has made. Likewise, behaviorists like Pavlov, Watson, and Skinner made this argument in the early to mid-1900s, although the field of psychology has largely rejected it now.

I would challenge the claim that you cannot objectively measure subjective experience because it has been shown that subjective experience (at least as measured by self-report scales) correlates with objectively observable phenomena like behavior and brain activity. If the variability measured by self-report scales of subjective experience were entirely random, unreliable, and erroneous, it could not consistently covary with objectively observable (or indeed any other) phenomena. Eventually, we may be able to understand someone's mental state with 100% accuracy just by observing the activity in their brains.