Why I am not a utilitarian
fakenous.net
fakenous.net
Deonological failures present themselves as conflicts of duty or moral dilemmas. It's not difficult to contrive these ourselves or easily find a ready-made list. To that end, the author's conclusion of defaulting to deontology is logically inconsistent - it merely avoids them by not bringing them up at at all.
I'm not sure I see the argument of the article that way (even though the author does end by stating that we should default to deontology). I would phrase the core of the argument this way: utilitarianism is a form of deontology! Utilitarianism ultimately rests on the ethical intuition that maximizing pleasure and minimizing pain is the only value that should be pursued. That's not an alternative to deontology: that is a deontological claim. And, once it's put baldly like that, it's a pretty questionable claim: is maximizing pleasure and minimizing pain really the only value that should be pursued? The article is simply posing scenarios in which some other deontological value conflicts with that value.
Systems of morals can all be expressed as rules, outcomes, or goals, so each can be phrased to subsume the others.
> Utilitarianism ultimately rests on the ethical intuition that maximizing pleasure and minimizing pain is the only value that should be pursued.
That sounds more like hedonism. Utilitarians don't necessarily believe that pleasure is perfectly aligned with utility.
Then say "maximizing utility" in place of "maximizing pleasure and minimizing pain". It doesn't change the fact that utilitarianism says there is just one value worth pursuing, whereas other deontological systems say there are multiple values worth pursuing. The latter seems to me to be far more in line with reality.
Here's the thing: one of the biggest problems to deontological approaches, as the parent pointed out, is that you end up with conflicts between values. But it's worse than that, because there's not a clear way to resolve these conflicts, and in my experience there's just a "hands throwing up" moment, along with a denial that the conflict existed, or an appeal to "maintaining values" without even recognizing that another value was abandoned.
So you could turn this on its head, and argue that if deontological and utilitarian approaches are equivalent at some level, why not advocate for the approach that at least offers a framework for resolving conflicts in theory? If two values confict, how do you decide? More importantly, how do you decide without appealing to utility at some level, including implicitly?
This gets into very high-level ethical reasoning issues and I'm not sure I would ascribe as purely utilitarian or deontological perspectives. But I think if you take the perspective of equivalence (and I think that's a reasonable perspective), there's not much grounds for saying you're not deontological or not utilitarian.
I'm not sure there is any general rule for this. That's why there are conflicts like this in the first place: because there isn't one consistent set of ethical rules that can be applied to every situation.
> More importantly, how do you decide without appealing to utility at some level, including implicitly?
"Utility" in this context is just another way of saying "what people value". But if there were one single scale on which we could measure everything that people value, there wouldn't be ethical conflicts. The whole point is that there isn't any such scale: people have multiple different, incommensurable values, and there is no general rule for how to resolve conflicts between them.
I have no problem with this, but utilitarians claim that utilitarianism is an alternative to deontology. If your statement here is true, then utilitarians are wrong in that claim.
It is if it recognizes that there is no magic framing that escapes moral conflicts. Utilitarianism claims to be such a framing, so showing that it doesn't actually deliver those goods seems to me to be a contribution worth making.
There are many cases where the "ethical" course of action makes many people miserable and where a simple, pragmatic, utilitarian solution would be beneficial. You don't have to go to extreme examples like the trolley problem or shooting down planes with terrorists in them. Just look at organ donations, for example. Here in Germany, it is opt-in. Why can't we just make it opt-out, or like in the US: when you get a drivers license, you have to decide yes or no?
I would consider myself somwhat utilitarian, but I have no problem reconciling most of the mentioned problems with common sense. Sometimes, it is by questioning the set-up:
> b. Framing the innocent > You’re the sheriff in a town where people are upset about a recent crime. If no one is punished, there will be riots. You can’t find the real criminal. Should you frame an innocent person, causing him to be unjustly punished, thus preventing the greater harm that would be caused by the riots?
The real problem here is that people are going to riot if their sense of justice is not satisfied. You're not just a sheriff but also a philosopher - the solution is to change society!
Most of the problems are variations of the classic "one person suffers greatly so many can have shallow pleasure". I don't know anybody who really proposes integrating over utility without any weight like this. Instead, most people would weight the injustice much greater than the pleasure. I maintain for example, we could radically reduce CO2 emissions and redistribute wealth to reduce poverty if we were willing to accept a couple percent GDP degrowth. That means on average, you might have to wait 2-3 years to get a new phone or a new car, and you might have to take a less than ideal job for a couple of years. But your inconvenience would be more than compensated by the increased utility for all.
Smith referred to him lovingly as "the never-to-be-forgotten Hutcheson." Oops.
Hutcheson also devised a system of "moral computation" (complete with a series of equations—and applicable to AI systems) and specifically indicated how it should be used in conjunction with human intuitions.
In the end, it doesn't much matter who did what, as long as good ideas got popularized. Bentham made an enormous contribution to developing and spreading utilitarian thought, and should be praised as such.
Thank you for sharing about Hutcheson - I've not heard of him before.
One of the best resources for philosophy reading is the Stanford Encyclopedia of Philosophy. Here it seems Hutcheson is credited with "proto-utilitarian" views / "anticipation of some strains of utilitarian thought".
His work on the inner sense of harmony and beauty is great psychology. Big influence on smith, whose first book was on Sympathy.
Oh, he also coined the term "inalienable rights" and wrote the text on moral philosophy that was used to educate over 60 percent of the signers of the declaration of independence.
Also had a big connection to secret societies and the whole western esoteric deal. Part of why Adam Smith burned all his papers (one may presume).
I really don’t think these are examples of utilitarianism.
So, one is comparing “a person dying” and “5 people not dying”, unless there’s something about the experience of the death-due-to-murder and the death-due-to-illness that differs and is taking into account as part of the outcome.
If you assign value to acts, then you don’t really have utilitarianism anymore, and you can probably describe nearly any ethical system in that language?
Maybe it's the fact that the act would have a major negative effect on the surgeon, or that admitting a justification for murder other than self-defense could open the door to abuse causing widespread harm. Ie, maybe our intuition is basically utilitarian but is constrained by rights and rules to protect against risks that are hard to quantify.
Many of these examples fail to take overall utility into account. The sadistic pleasure is another prime example. You must consider not just those who would enjoy the torture but also those that would be repulsed by it.
Same with the cookie. Rewarding a serial killer would make much of society unhappy. From a utilitarian point of view, giving the cookie to Mother Teresa (generally rewarding those who act kindly to others), is a very straightforward choice. The individual perception of cookie sweetness contributes very little to the overall utility.
It is in fact the case that we do often do bad things for more important goals. We kill people in wars. We support legal systems that we know sometimes imprison innocent people. The government takes money from people against their will for their own purposes, which isn't stealing mostly only because it's written in law.
What's the difference? Well, in those latter cases, it is actually considered better in sum for that action to be taken. For most of the examples in the post, there are simple and clear arguments that it is not better in sum, and the only reason the article is claiming otherwise is that the author is pretending to be dumber than they are by not mentioning the rest. So don't do that.
(And for some cases, though you wouldn't know it from the article, the full utilitarian calculus does actually say that uncomfortable things are the right thing to do, and in those cases you should just bite the damn bullet.)
This is a really good point and one that doesn’t get raised enough.
For example, the murdering doctor example assumes that there are no dangerous second order consequences to randomly murdering people. That isn't a safe assumption. A doctor might reasonably conclude there is enough value in having trustworthy institutions that the locally optimum choice is different from the globally optimum choice after evaluating the risks.
A utilitarian could make either choice in any of those examples if they disagree with the assumptions implicit in the question - which isn't really allowed in this sort of thought experiment. The situation is artificial because the author is allowed to create certainty and constraints which don't map to the reality where these problems will be encountered. Any moral framework will give weird results if applied to artificial situations, so the construction of artificial situations where the author objects the utilitarian answer is unless as evidence.
The aversions we have to this kind of inhumane utility juggling are because we instinctually recongize the second order and knock-on effects without needing to specify them. Dehumanization, inequality, precedence - these are all utility-real things which a responsible utilitarian incorporates into their utility calculus as major negative impactors.
When you strip the situation of any practical relevance, of course the conclusions are 'stupid' because the criteria you apply to determine that stupidity are based on a practical reality.
Someone being a "stupid utilitarian" a subjective judgement.
> then you are NOT actually summing up all the resulting negative utility in your so-called-utilitarian calculations
Well that's the trick isn't it. What's the correct calculation? There is no objective answer to this. You can change the calculation to suit your opinion.
Consider: it could be entirely consistent for a Utilitarian to make the utility justified decision that utility calculus is outside of their capability, and a Deontological framework would serve as their best in-situ heuristic. They would still be a Utilitarian, even thought their black-box criteria resembles a deontology.
What you're not allowed to do is say, as the article did, that X has the highest utility when strawmanning a utilitarian, only to turn around and say X sucks once the strawman is over.
Alternatively, on a normative level, you can still debate the merits of the framework when compared to other systems (deontological - virtue - care - etc.) by beginning with meta-ethical axioms and resolving them to normative positions. Consider an ontological framework which doesn't recognize free will of others - this could dramatically inform a delimma between a utilitarian system and an egoist one.
I'd be more open to questions about how a rational person might figure out what the best things to do are in the face of uncertainty, if the article's examples didn't include things like should we torture Jewish people for the enjoyment of Nazis? The right intervention to that is much more baseline, of actually stopping to think about what one is saying.
I may not be a utilitarian, but I do think war is a bad thing, so I’m not sure how this is a great example.
This is kind of missing the point. Even if it were to result in a "better world", the intuition is that it is still wrong to murder the innocent man to save 5, just prima facie. This is a problem for the utilitarian.
Their best response might be to just bite the bullet, but it significantly weakens their argument overall.
I disagree. To me it seems fairly obvious that it is morally correct to murder one innocent person to save 5 innocent people. Not only that, but we often as a society make this exact decision (e.g. send young people to fight in armies, send firefighters to rescue people despite potential harm to themselves, etc).
I don't know why this supposedly weakens utilitarianism - the whole point of the philosophy is exactly to point out that this is the morally correct choice.
The problem with this and the other examples in the article is that there's a big difference between the hypothetical examples often given, which assume no other effects, and the real world. In the real world, we often only allow these kinds of decisions in society-wide settings - e.g. the examples I gave above (armies, firefighters). We don't allow people to just choose on a case-by-case basis who should live or die, because it would cause far more problems (and thus, negative utility) to do so.
Yes, if you only care about pleasure, you will make horrible moral choices.
Now if we include in our definition of goodness things are are actually important like survival, stability of order, justice and fairness, scientific and personal progress and so on, we might get more practical. Of course then these decision become multi-dimensional and and highly dependent on your personal priorities but that is the real world for you.
They are comparable to to the question "But if you want to live a happy life, why not inject yourself with heroin and feel super happy for a few hours? It would be the best way to live a happy life today."
Upholding personal rights, predictability and principles will maximise human happiness in the long term, while optimising for short-term happiness will often cause suffering in the long term.
Some examples:
> Say you’re a surgeon. You have 5 patients who need organ transplants, plus 1 healthy patient who is compatible with the other 5. Should you murder the healthy patient so you can distribute his organs, thus saving 5 lives?
A society where people are deathly afraid of sending loved ones to the hospital, for fear they might be killed and used for organ harvesting, will lead to the collapse of healthcare and cause far more suffering than temporarily increasing the supply of organs will prevent.
> You’re the sheriff in a town where people are upset about a recent crime. If no one is punished, there will be riots. You can’t find the real criminal. Should you frame an innocent person, causing him to be unjustly punished, thus preventing the greater harm that would be caused by the riots?
If this framing of the innocent is discovered, which is not unlikely, it will lead to complete distrust in law enforcement, and over time the corruption of law enforcement will cause far more suffering than stopping this specific riot will prevent. And the unjust punishment might cause riots for a different reason, anyway.
This might even match 1-to-1 with your intuition, if your intuition is "This seems logical on paper, but I don't want to live in a society where this happens, so I'm against it."
TL;DR: We live in a society.
> pleasure/enjoyment (and the absence of pain), > Say you’re a surgeon. You have 5 patients who need organ transplants, plus 1 healthy patient who is compatible with the other 5. Should you murder the healthy patient so you can distribute his organs, thus saving 5 lives?
How is murdering a guy, not painful for the guy? So I guess that's clear how utilitarianism handles that
I think if we consider the frustration and pain there is nothing that is not clear
The examples are simply absurd. As many other commenters have pointed out, he simply doesn’t evaluate them reasonably.
You can certainly make the case that this is a real problem with utilitarianism - i.e. that there is no good function for evaluating scenarios.
But that’s not what the author s arguing. He is arguing that utilitarianism leads to repugnant trade-offs.
He’s not even wrong. He just doesn’t seem to have thought deeply about it at all.
This simulation was indeed very simplistic and cannot represent or predict reality, but neither can you. Until you test it, you won’t figure out if it’s beneficial or not to someone, everyone, anyone. Watching these dilemmas in isolation does nothing to your understanding, whichever philosophy you use.
We don’t build our way to our goals, we explore the insanely complex network of possibilities and choose (and stick to) those which worked good enough. It’s like an edge of a fractal, sometimes a little change converges to nothing different, sometimes it turns everything upside down.
> If you don’t accept ethical intuition as a source of justified belief, then you have no reason for thinking that enjoyment is better than suffering, that satisfying desires is better than frustrating them, that we should produce more good rather than less, or that we should care about anyone other than ourselves.
We have to take the author's word for this, because he does not offer any support for his claim. I wonder whether he believes that his personal ethical intuition can/may/should be generalised. The structural similarity to "whosoever does not believe in (my) god has no base for ethics" should serve as a warning.
> If you do accept ethical intuition, then at least prima facie, you should accept each of the above examples as counter-examples to utilitarianism. Since there are so many counter-examples, and the intuitions about these examples are strong and widespread, it’s hard to see how utilitarianism could be justified overall.
The auther is willfully ignorant about utilitarism and should make amends.
Organ harvesting: How bad is the fear in every healthy person for risking to be harvested? Healthy people are much more that people in need organ transplants. How that fear condition their life. Is range generated? With what effects on in need organ transplants?
Framing the innocent: If the action from sheriff comes out and don't get punished what's the effect in people that that fear to be framed for no reason? What's the effect for people losing trust in justice? If you need just a temporary guilty just fake it as long as you find the real criminal or you improve the situation with other means. Obviously situations abstracted from the context and for others lines of actions lead to absurd solutions.
Deathbed promise: Considering the aftereffects of this case are so complex that giving a clear response is quite complex at best.
Sports match: Just save the person and let the other people know that they helped to save a person in need. Gratification will be much more that annoyance for short interruption and memories for that extraordinary events will be good too. If the person was just experiencing a little distress for the latest minutes of mood landing just let him experience it. Utilitarianism let you balance different situations.
Cookie: If we are talking about cookies who care. If we are talking about reward we can balance two things: - a gentle act toward a mad person can have much bigger effects - letting people know that good people get rewarded can generate more good people and more good with time
The Professor and the Serial Killer: Professor do good mainly with his work and with his personal life. Charity is only a small part. How being a lawyer would affect this total outcome? These more money he earn come from somewhere. What's the total effect?
Excess altruism: An altruist get joy from helping others so giving the cookie to Sue the total pleasure is maximized.
>Say you’re a surgeon. You have 5 patients who need organ transplants, plus 1 healthy patient who is compatible with the other 5. Should you murder the healthy patient so you can distribute his organs, thus saving 5 lives?
There is no mention about possible complications with surgery (all five patients could die shortly after recovering from surgery) or whether there are other humans in the world that could donate the organs without sacrificing a life.
In real life you would wait (i.e., do nothing, do no harm) to find alternatives, which would redefine the problem. In the theoretical world there's no gain in waiting because the problem is strictly defined and immutable.
Also, if you have been in the software industry long enough, you have probably been in a situation where you had to deliver buggy or incomplete software because your customer wanted it yesterday.
Personally, I'd rather let someone know the stuff they are getting is not as good as it should be (yet) than fail to meet a deadline and permanently damage our relationship.
This is of course heavily dependent on the problem domain. The failure of a seismometer and a hypercasual puzzle game have very different consequences.
This is not the case, if it would be philosophy wouldn't exist instead we would just have people un-necessarily explicating their intuitions, like the article.
1. Utilitarianism asks me to choose to save 10 strangers over my own kid, as this would maximize human happiness.
2. The bond a parent shares with their kid is among the most foundational examples of human happiness.
3. If human happiness is so flimsy that I should be willing to destroy the root of my own in the advancement some abstract ethical theory, why is human happiness worth maximizing in first the place?
My understanding of utilitarianism is that other factors than keeping someone alive should be considered. Is it utilitarian to save 10 elderly people over the life of a child? It might not be, depending on how long the elderly are expected to live, and if the remainder of their lives is as valuable as those of a child (eg, will their remaining years be unproductive and sickly due to age, while the child’s life may be full of productivity and health).
It does not make that assumption.
Also, if that decision were made, it's not in the service of advancing some theory, it's in sacrifice. Wouldn't to do otherwise be a selfish decision? And if not, what does it say about maximizing happiness? Whose happiness is relevant? The 12 people involved, or whoever might be privy to the decision?
Some of these things might argue against utilitarianism, but I'm not sure any other ethical framework would be more immune to criticisms.
Sometimes I think these kinds of decision scenarios are fundamentally intractable, and don't really speak to reasons for adopting or not adopting any particular ethical paradigm.
In theory no, but in practice it is the concrete instantiation of utility most commonly proposed. Moreover, I think any definition of utility we invoke will be susceptible to a similar critique.
> another significant problem is in quantifying happiness
I agree with this, and I think the measurement problem is insurmountable. But I still think Williams' criticism is deeper. If we magically solved out the measurement problem (which, again, I think is impossible - the question is ill-posed) I would still be opposed to utilitarianism, because whatever notion of goodness we invoke becomes trivialised when we boil it down to a liquid currency.
> Sometimes I think these kinds of decision scenarios are fundamentally intractable, and don't really speak to reasons for adopting or not adopting any particular ethical paradigm.
I think they can be useful, but analytic philosophers have a bad tendency to lean far too heavily on them. Often results in missing the forest for the trees.
I think it is good to point out the flaws in a short and concise way. An achievement, really. It is more difficult to be short and concise than to write long articles.
Then things like enjoing one cookie slightly more than somebody else isn't a problem, because enjoyment of ten cookies is not much greater than enjoyment of nine cookies and you can have a spare one for somebody else.
Really? You picked the easiest example in the article. Try a more interesting one, like, say, organ harvesting: how does logarithmic utility deal with that?
What is the second order effect in this case? "Humans might be killed for organ harvesting. This leads to widespread uncertainty over who would be killed next. And uncertainty leads to other issues."
For the organ harvesting scenario, if you were behind the veil and had to choose what society you'd enter, would you choose to go to the one where a yearly physical might end up with you dead while you're parted out for others in the hospital—and also might mean you get your organ replacement lickety-split should you need one—or would you rather go to the one where this does not happen?
When it comes to "real" case: it causes distress to the whole population and transplanted organs have consequences (e.g. immunosupressants) that lower their value.
In any case, your claim that I responded to was that "most, if not all, of these problems can be done away with" by adopting logarithmic instead of linear utility. You haven't said anything about how logarithmic utility "does away with" the problem of deciding whether organ harvesting is ethical. Or about any other of the scenarios in the article except the cookie one. So it looks to me like you are backing off from your original claim.
Many people are utilitarian to some degree, and most – including those who might self-identify as utilitarian – would say there are limits to it, too. There are extremists who push the utilitarian limits very far, but those are relatively rare.
The same applies to many other philosophies. Many people – on both the left and right – would agree that anyone "should have the freedom as they want, without too much interference of government or other organisations", and are libertarians to some degree. It's only when applied to the extremes without critically considering the trade-offs that it becomes silly, or even evil.
The author prefers deontology; that okay, but it's not hard to come up with crazy examples for that, too. If a patient's life can be saved by investing all the hospital's financial resources then deontology, pushed to its extreme, says we should do this. This would be silly, and no doubt the author would agree, because you don't have infinite money and need to think about the patients that come in tomorrow, too.
In short, I feel this article is rather myopic, as well as an example of "nut picking" (that is, pick the nuttiest/craziest views and then criticise the entire concept with those views).
If you think killing the healthy man is wrong, you probably don't agree with utilitarianism.
The only thing that this text points to is that it's better not to be strict utilitarian if you have iq below 50 or a mentality of evil djinni.
> There is a large number of Nazis who would enjoy seeing an innocent Jewish
> person tortured – so many that their total pleasure would be greater than the
> victim’s suffering. Should you torture an innocent Jewish person so you can
> give pleasure to all these Nazis?
This makes sense if the weight of pleasure is uncapped, but imagine a utility function where lots of pleasure still only has a bit of utility (say 10) whereas very extreme displeasure could have much more negative or nearly uncapped negative utility (say, -(10^20) for torture). Then no number of Nazis that could ever physically exist would be able to cause this situation to have negative utility.
Thinking in those terms, I find the criticism weak.
Just like most ethical conundrums, these basically all rely on carefully controlling the bounds of the hypothetical to be both short-sighted and unrealistic to the point of farcical. How does the sheriff know for certain there will be riots if he doesn't hang an innocent person? He doesn't. How will all those sports fans feel if they learn that an innocent man needlessly died to avoid interrupting their feed for 10 seconds? They would probably feel terrible. Are we really just going to accept these extremely unrealistic hypotheticals where the author assures us his two options and consequences are genuinely the only possibilities? How is the surgeon going to feel if he murders someone as part of his job? It wouldn't take long for him to burn out and commit suicide, adding another needless death and depriving the world of a presumably talented surgeon.
In general most of the examples seem to imply that a utilitarian would do the very short-sighted option that has the biggest immediate increase in pleasure, while completely ignoring all the other negative long-term effects of their decision.
The irony.
Many of the examples in the article are constructed to omit outside observers. In the real world, this is very unrealistic. Our intuitions about the importance of acting consistently reflect the risk that we'll be discovered doing something abhorrent. Reputation is important.
In fact, if we make the examples more abstract and impersonal, then it's much easier to see how utilitarianism makes sense. What if there was no actor making a choice; instead, the two possible outcomes happen by chance or natural causes. Say the healthy patient dies from a freak accident, allowing other patients to be saved. This seems like a better outcome than five people dying, all else equal. It's only the possible effect on the doctor's reputation that makes the original a moral dilemma.
Similarly, in the riot example, if the innocent person was implicated by accident, that would be a better outcome than many innocent people dying to the mob. It's only introducing the sheriff and his reputation that makes it a dilemma.
This question is constantly trotted out and it's a great example of the deeply unreal world ethicists insist on operating in. In the real world, you don't, because making people people live in a world where that happens costs more than doing it would benefit you. Doing the murder is therefore not, I repeat, not the utilitarian answer.
They always refuse to pose the question in a way that might actually make students consider whether the murder is worth it. For example, suppose the seven people are the crew of a spacecraft sent to divert an asteroid from wiping out human civilization. Is murdering one to save billions justified?
No.
In fact, we make far worse moral calculations than that frequently. Eg, if you've ever supported a war, you've agreed with killing completely innocent children, often in horrible ways, for the sake of whatever benefit the war might have had for the side you support. And there have been plenty wars fought for reasons other than self-defense. Take Iraq and Afghanistan for instance.
A nucleur missle is about to fire off and kill everybody in your country. You can only shoot it before it launches, missle defense can't catch it after launch.
You can preemtively destroy the missle, but an innocent child playing near the missle silo is going to be killed.
Do you stop the missle? The child's death will be called collateral damage. Nobody was "murdered".
[1] For simplicity ignoring all the complications of who wants to attack, probably the government or the military, and who would be victim of the prevention, probably military and civilians, and that you therefore might have to kill people that do not support or even oppose the attack.