I will argue that John Stuart Mill's greatest happiness principle (GHP) should be revised to avoid problematic implications and to better fit humans' intuitive sense of morality. Furthermore, the resulting revised version of GHP is actually quantitatively (and thus qualitatively) equivalent to Rawls's theory of justice as fairness.
My question: did anyone make this argument or a similar one reconciling utilitarianism and Rawlsianism (if that's what you call it)? Also, am I misinterpreting Mill's or Rawls's theories or making mistakes in my reasoning?
Sorry for how long this turned out (it seemed a lot shorter in my head).
John Stuart Mill's greatest happiness principle (GHP) states that "actions are right in proportion as they tend to promote [the indiscriminate sum total of individual] happiness." While Mill does not explicitly state the bracketed bit in his description of GHP, he and his followers seem to assume it in their application of GHP. However, his argument does not justify that bracketed bit, and I will argue that it is actually the source of many of utilitarianism's problems. I will revise GHP to deal with these problems and come up with a BHP, or best happiness principle.
Problem 1: inequality is bad, but Mill's GHP does not seem to take that into account. Indeed, when allocating happiness in a society of 10 people, Mill would be indifferent between giving 1 person 11 units and 1 unit each to the other 9, and giving everyone 2 units (they come out to the same sum). Intuitively, this is wrong; we feel that a more equal society is better, ceteris paribus. So how do we adjust this? We employ the principle of marginal utility. If one person is already very happy, giving them another unit of happiness should not be as valuable as giving a very sad person happiness (intuitively, it is more morally valuable to cheer someone up when they are sad than it is to stoke a happy person's ego). (A classic example of diminishing marginal utility: it is more valuable to give a starving person some food than to give that food to someone who has just eaten their fill at an all-you-can-eat buffet.) Therefore, we can apply a utility function to the happiness of each person in the society with this in mind (making it increasing with a diminishing slope), such as the square root (sqrt) function. We sum up all of the utilities. We should then square the total to get back to happiness units. When we apply this to our previous example, we have (sqrt(12)+9)^2=155.35 for the unequal society and (sqrt(2)*10)^2=200 for the equal society. Thus, we have accounted for the immorality of inequality.
Another problem we run into with utilitarianism is when we apply it to situations of changing the population size. Is it better to let overpopulation run wild, exhausting the earth's resources and making everyone else slightly less happy? Original GHP as well as our current modification of GHP would say yes; the sum total of happiness would be increasing. This is intuitively wrong, however. It seems that we want to maximize people's quality of lives (or the average happiness), rather than the sum total of happiness. GHP (in its original form and in our modified form) yields the “repugnant conclusion” that a very large society of miserable beings is better than a small society of happy beings (Parfit). Thus we can improve on our above modification of GHP by taking the average of the sum of the utilities of each person's happiness (getting the average utility), and then squaring that to convert to happiness units (this is analogous to going from total utilitarianism to average utilitarianism). If we let h_i be the happiness of the ith person in the world (assuming everyone in the world is numbered), where i is between 1 and p (p being the total number of people in the world), we are maximizing for: ((sum from i=1 to p of sqrt(h_i))/p)^2. I'll refer to the principle that humans should aim to maximize this quantity as the best happiness principle (BHP).
There are two other comments (not necessarily modifications, but rather corrections of anticipated misinterpretations) I should make on this principle. First, one objector might say that BHP has the repugnant conclusion that we would want to kill people who are less sad than average. This is certainly not true; killing anyone would violate their rights (which exist because human rights are necessary for human happiness, as Mill points out). Another way to look at it: killing someone for this reason would cause everyone else to become terrified that they or their loved ones would be next on the utilitarian chopping block, significantly reducing inequality-adjusted average happiness and violating BHP. So this is not actually a sound objection to BHP. Another objector might say that BHP would have the repugnant conclusion that a society in which one person would have infinite happiness and everyone else is as sad as can be would always be better than a society in which everyone had some happiness. This ridiculousness of this objection stems from the fact that no one can be infinitely happy. There is a limit to how much happiness one person can have. So this is also not a sound objection. One last possible objection: BHP doesn't take into account the moral value of nature. It actually does; killing animals or harming nature would be immoral because animals and nature make humans happy. One might criticize this as too anthropocentric, as it bases the value of nature in terms of how it makes humans feel. This may be a fair point, but I will fight back by pointing out that I, for one, would have no problem seeing mosquitoes eliminated from the earth, if it weren't the case that such an extinction would hugely disrupt many ecosystems. Lastly: I do not intend BHP to be a rule-of-thumb for people to use in their everyday lives (it seems unlikely that someone could make each of their decisions to maximize inequality-adjusted average happiness). Rather, it should be a general goal for humans as a whole, which we should take into account in the design of social institutions (stay tuned), as well as secondary moral principles.
Now, let's consider Rawls's 'original position,' the societal staging ground where people haven't yet discovered their natural resources/attributes (intellect, work-ethic, race, gender, etc.). All people in the OP are rational and mutually disinterested (as he puts it). We want to structure society such that these people would want to be in our version of society as much as possible. That way, although people cannot consent to entering our society as babies, we have the next best thing: ensuring that, if babies could decide whether or not to enter our society as rational beings, they would do so (because there would be no different structure of society that they would prefer). Rawls is assuming that an optimal society is always preferable to having no society (I think this is a safe assumption to make in our modern world). Assuming that any given member of the OP ultimately wants to maximize their own happiness (since they are mutually disinterested), how should we structure society such that they will want to be in the society as much as possible? One might say that we should maximize the expected value of happiness (this is equivalent to the average happiness, since our OP member has 1/p chance of being any particular person in the society, so the expected value is the sum of each individual's happiness all divided by p). If we did so, their expected situation would be as good as possible. However, as decision theorists would attest to, expected value is not all that matters in evaluating a probabilistic situation. Due to the diminishing marginal utility of goods, people are generally risk-averse in such situations. I will not justify this beyond saying that it is well supported by decision theory and empirical data. Thus, a rational member of the OP would not want to maximize the expected value of their happiness in the society, they would want to maximize the certainty equivalent of their happiness in the society. The certainty equivalent of a probabilistic situation is the amount of the good in question such that the actor would be indifferent between taking that amount with 100% certainty and going through with the probabilistic situation. For example, if someone with a sqrt utility function was offered a 100$/hr salary if a flipped coin landed heads and 0$/hr if tails, the certainty equivalent would be (sqrt(10)/2+sqrt(0)/2)^2=25. This is lower than the expected value of 50$ (by 25$) because this person would be willing to pay to mitigate the risk of having to work for no salary (this concept explains how insurance companies make money). You might see where this is going. Our OP member would want to maximize the certainty equivalent of their happiness in the society, which can be expressed as: ((sum from i=1 to p of (1/p)*sqrt(h_i)))^2=((sum from i=1 to p of sqrt(h_i))/p)^2, which is the same quantity as that which BHP states should be maximized. And thus our modified utilitarian theory is reconciled with Rawls's concept of the original position.