Assume, for the sake of argument, that (1) human agents make decisions; (2) when human agents make decisions, they choose in such a way as to maximize expected utility; (3) all utilities are directly and accurately known at all times. The first assumption is relatively unproblematic from a naive point of view. Even for a hard determinist, though, the first assumption should be acceptable because no individual is a perfect predictor. From the point of view of any specific human observer, all other human agents are sufficiently unpredictable that some process that approximates to "decision-making" can be assumed to transpire. The second assumption is a grounding assumption of utility theory. I will not attempt to justify it. If you wish to argue with it, however, please feel free. The third assumption is obviously the least plausible, and I will not pretend it is anything more than science fiction. I would request that objections to it be left aside for present purposes, however.

One version of utilitarianism might adopt the principle that the most socially desired outcome is in all cases that which maximizes aggregate utility, regardless of which human agents receive utility benefits, and regardless of whether any suffer disutility. A concept such as "rights," or any similar deontological concept, is inadmissible on this view. Rather, regardless of whether an actual outcome causes even extreme disutility to a large class of human agents, if this is the outcome that maximizes aggregate utility, it is the most socially desired.

One can make this view plausible by adding assumptions about how most people's utility functions actually work: that they do not tend to profit from using violence against others, for example, and so it is very unlikely that extreme discord or violence will result. In addition, most people have evolved in such a way as not to receive pleasure from the suffering of others. Defenders of the free market on this forum, including myself, often advance these very arguments.

With this view in mind, imagine a 2 person society consisting of Person A and Person B. Person A, who has some fairly sadistic tendencies, decides to punch Person B in the arm quite hard, causing pain and injury. Person A receives +4 utility from this act; Person B suffers -2. Summing the utilities, the utilitarian view I described above would state that the socially desired outcome in this scenario is for Person A to punch Person B. In fact, this will be true until the disutility caused by the pain and injury from the punching is greater than the utility received from the sadistic satisfaction.

How do people interpret this utilitarian position? If the assumptions above are satisfied, are you comfortable endorsing the view advanced above, i.e. that the most socially desired outcome is that which maximizes aggregate utility? Would you endorse some further rule regarding the rights of the people involved to be free from disutility caused by others? Would you restrict your claim regarding disutility to disutility caused by violence or coercion? If so, why?

Most importantly, if you advance a contrary moral argument with an affirmative content - one that states a rule regarding what is right or best - why should your claim be entitled to moral authority over any number of contrary, competing claims?

we need to have like a political spectrum next to each poaster in this forum so i can figure out where people are coming from

Didn't read it all, but you can't sum utility as you assumed.

Utility theory maintains that I could multiply my utility function by any positive number, and it's still a valid utility function (and will act in exactly the same way, with simple decisions and also decisions over lotteries). Because of that, comparing something that's utility -4 to me to something that's +2 to you is meaningless. I could add 6 to my utility function and then multiply that function by 3, and now that action that was -4 for me is now +6.

The example I have above would be asking 'who gets a certain good', and we'll give it to the person who has the highest gain in utility from that. So with one utility for me, the good would go to you, but with the other utility function, I get the good. Even though nothing has changed.

this is an example of why its occasionally useful to read poasts

You might just about be able to conceive of situations in which you could justify it in the short run, but that is not the only thing which is to be taken into consideration. It's pretty hard to argue that satisfying sadistic desires would ever maximize the utility of a society in the long run. Any society which found such behaviors morally permissible (even good, as an increase in net utility is always moral) would soon experience massive losses in utility from the lack of security people felt, to the encouragement and thus increase of sadistic behavior. Taking isolate incidents like this cannot really inform us about a moral theory which has implications for a whole society. Even if we could isolate it, I believe there would always be a net utility loss due to the above reasons - loss of security and increase of sadistic behavior.

You might also want to consider Mill's concept of higher pleasures as a refutation of hedonistic strains of utilitarianism. Sadistic desires would, to Mill, be of the basest and most stupefying pleasures, as such the pleasure they bring does not have the same weight as the 'higher pleasures.' Although the higher pleasures usually refer to cultural pursuits, I believe Mill would consider the utility of the security of Person B, greater than the utility of Person A.

Your example is a slightly tweaked 'sacrifice' scenario, which almost always ignores the huge importance placed on security in utilitarian theory.

I am a fairly strong utilitarian.....but you have to ask what utilitarianism is being used for.
Typically, it is used in the context of how we should organize society, and what we should collectively accept, encourage, and prohibit. In this context, you need to design rules that can clearly be followed, even when you don't know the utility functions of everyone involved.

So I guess my answer is that, within the hypothetical situations using the assumptions that you make, the action of A punching B would be permissible.
But you could never write a system of law that would permit it, unless you felt in general that (random person X) punching (random person Y) provided a net utility benefit (or could in some other way easily isolate the situations where this would be true).

How do you feel about the assertion that rule utilitarianism often breaks down relatively easily into act utilitarianism?

I'd argue that utilitarianism does not just concern political and legal systems - it is also covers a system of personal morality. So taking the politics and law out of this situation is fairly justified, treating it in a complete vacuum is not though.

You mean when the OP claims that a good decision making method is "With this view in mind, imagine a 2 person society consisting of Person A and Person B. Person A, who has some fairly sadistic tendencies, decides to punch Person B in the arm quite hard, causing pain and injury. Person A receives +4 utility from this act; Person B suffers -2. Summing the utilities, the utilitarian view I described above would state that the socially desired outcome in this scenario is for Person A to punch Person B. In fact, this will be true until the disutility caused by the pain and injury from the punching is greater than the utility received from the sadistic satisfaction." Or earlier when he talks about summing an entire society's utility. You just can't do that.

My post says exactly why this isn't a good metric. Thus, to clearly answer one of the OPs question, no I am not comfortable endorsing this view (or however the OP phrased it). Basically when you start comparing one person's utilities with another's, it's a complete non-starter.

Oh, and if you want to make decisions by comparing utility, please add 'bout three fiddy and multiply that by eleventy billion. kthx.

This is fairly easily to reconcile by supplying additional assumptions that render interpersonal comparison of utility meaningful. We can simply assume that all utilities occur in a common, equally-sized unit. I consider that assumption an ancillary assumption necessary to make my third assumption above meaningful. Indeed, I intended it to be included in the original set of assumptions, because reaching the question I am interested in requires that it be supplied, but you are right that I did not explicitly state it. For the rest of the discussion, let us assume that 1 utile for Person A = 1 utile for Person B.

But yeah, the problem of aggregating utilities is one I have yet to see utilitarians give a convincing refutation of. John Rawl's A Theory of Justice well and truly put the boot into this issue with utilitarianism, and I'd highly recommend it to you DrModern if you haven't read it already for a lucid and convincing argument against utilitarianism. If you have read it, I'd be interested to hear your thoughts.

I guess the way I feel about it is that I believe it is not true.
How exactly do you write a law that permits A to punch B only when it is a net utility gain (given that no one can know or prove utilities of the individuals involved)?

As for personal morality....I guess it is moral for A to punch B if he is certain that it is a net utility good (and he can't just consider A and B in the equation). But from a legal standpoint, he should go to jail regardless.

Well, if you want to start doing that, you'd have to do some sort of normalization--since just setting 1 util for A = 1 util for B leads to problems. Maybe I'm just easy to amuse, and so everything I do gives me more utility that you (or vice versa). Who's to say? *Maybe* normalizing utilities to be within certain ranges takes care of some/all of the problems I mentioned above, but there's still some skeletons lurking around.

By the way, if you changed your question from utility to willingness to pay, and instead of using the sum of utility but rather the sum of surplus, I'd get behind most things decided with that metric. That's concentrating on the efficiency of the market though, and I will reserve some equity considerations to be allowed into what I think ought to be done.

Note, with the willingness to pay (and associated aggregate surpluses), you are comparing my dollars with your dollars, and those can be directly compared. And I believe you can adjust your assumptions in your OP to deal with WTP and surpluses just like you did with utilities.

You're missing the moral theoretic issue with which I'm concerned by adding concern over normalization for utility. The salient consideration is that above, utility for Person A > utility for Person B, such that, on the view advanced, Person A punching Person B is the action endorsed as the socially desired outcome because it offers the highest aggregate utility. To many people, this is not a morally acceptable view. I am asking why they believe this to be so, or, on the contrary, if they would endorse this view in principle. In other words, if I supply sufficient assumptions within utility theory to make the hypothetical solvent, why would anyone object?

In addition, I don't think your concern for normalization is a genuine issue given the assumptions. 1 utile is in all ways equal to 1 utile. The "same" action can generate more utiles for a different person, so there is no problem of being "easily amused" or "difficult to please." Saying otherwise is stating a normative view.

I'm not really ignoring them, I'm simply putting them into a meaningful framework. Unless you're adding something into the assumptions that I disagree with, that somehow you can make your scale of utilities exactly equal to mine. It's sort of like saying 'one degree is equal to one degree' when comparing Fahrenheit to Celsius. You can try to normalize them further, but if the Fahrenheit is measuring the temp. in Southern California and Celsius is measuring something in the middle east, the normalization is also not going to work. This is the issue I was talking about--everyone has their own units for measuring utility (say F versus C), and you can't just assume away differences, and there are issues with normalizing. That's why I am saying the question lies on very shaky ground.

Now, you could ask the same question though with willingness to pay and surplus. If someone wants to hit me, and they're willing to pay $10 to do so, and I'm only willing to pay $4 not to get hit, then you have the same issue you did with your utility formulation. To that question, I would refer you to Coase's theorem. We will get the efficient outcome (assuming a small number of people, and property rights that are transferable). If I have the right not to be hit, then the person can pay $7, and we're both better off (and the efficient outcome occurs). If the person has the right to hit me, then I will not be willing to pay him enough to not want to hit me, so he will.

There may be other ways to reach the optimal allocation, and Coase's mechanism breaks down when there are a lot of participants, but that (I think) answers most of the questions about one person wanting to hit another one.

The question in this case then becomes who should be assigned which rights, and that's more an equity question than efficiency question.

This reminds me of the senior project I did for my international studies major on the utility decisions that go into violent vs nonviolent actions for terrorists. It was centered on Chechnya and Russia. My mentor was Russian (frequently back and forth and taught at my university and in Russia at various times) and she gave me a B because while getting more or less benefit from one action or another sounds good its really because 'they're born and bred to commit violence'.

I'm not sure I understand why punching someone in the arm just because he enjoys it gets him positive utility unless there are no consequences for such an action.

ShaneP,

If you don't want to play along with the hypothetical because you disagree with the assumptions, that's fine, but I think there are interesting questions about arguments for and against rights that you're missing as a result.

I did sort of get into the interesting rights questions (a bit at least) in the previous. I probably have to get some stuff done after the SD/NY game so I might not be around, but as I said, I think you can get the exact same questions without the troubles by using what I was suggesting. But one of the questions you asked (I'm paraphrasing I know) was whether you agreed with the procedure. I didn't, for the reasons I mentioned, but also gave you a way to formulate the question without those criticisms, while retaining those interesting questions. Hopefully when I come back later there's some discussion of those.

That's fair. I'm hoping this thread will generate some interesting theoretical discussion, as Politics in general seems to be in somewhat of a lull lately.

I have read it. I'd be happy to summarize my thoughts later in thread after some further discussion of the primary question has had a chance to develop.

I am totally comfortable with the idea that A hitting B is fine under the assumed situation. I'm not necessarily sure what relevance this has to the Real World.

Does B have no right not to be hit, even under the assumptions?

Add me to the camp which really isn't sure what you are doing when you're adding these utilities. I'm not sure how to combine societal welfare/happiness/utility in this way.

If it really makes you so happy, I might even let you punch me in the arm...

I guess I'm not sure what you consider your primary question, since on the one hand you say it is plausible to assume that people's utility functions will exist so that violence is never preferred, but your toy example includes an example of violence.

If there are two people, and the consequences of A punching B are such that A is happier and B is less happy (taking into account everything, e.g., the pleasure B receives from seeing A so happy), I just don't see how to construct the argument that we ought to have A punch B.

The question of "normalizing" the utilities to compare them without any other external reference truly seems insurmountable to me. If I went to person B and said "you know, if you cared more about not getting punched in the arm, we wouldn't let A punch you" I'm sure that B would reply "Care more compared to what?" or "I care a lot!".

The only way any of this would make sense is if there's something else that B cares about more that B is getting because A can punch B in the arm. The problem with toy examples with interpersonal utility is that there's no way to do this sort of trading, in the manner as ShaneP discussed, and so we end up with the question: "Should A get B's cookie?" being answered by "Yes, because A wants it more." (What does that mean?)

And even if that's a fine solution, I don't see how we get from there to a more general moral framework.

Explain? I'm not sure what the parameters are in this hypothetical.

Perhaps the shortcoming is with my articulation of the problem. What I am trying to illustrate is that certain utilitarian views, even those that are completely neutral with regard to whose welfare takes precedence, can generate social recommendations that show no regard for what many would regard as more or less inviolate rights - for example, a right to be free from initiation of physical attack by others. I am curious if champions of utility theory - practical considerations about the plausibility of the assumptions aside - are in principle prepared to embrace such a view in principle. If they are, have they abandoned any normative conception of rights? If they have a normative conception of rights, what is it, and how is it justified?

From my understanding of it, and its limited to two classes some 4 years ago mainly based on terrorism, A's utility is not affected by B's rights. If A is getting enough pleasure out of it to overcome the fine or jail or whatever negative consequence there would be it doesn't matter that B has a right to not be hit.