Tuesday, August 7, 2012

Utilitarianism, part 6: To do, or not to do

This is the sixth post in a series about utilitarianism.  For an introduction, see the first post.  For a look at total vs. average utilitarianism, see here.  For a discussion of act vs. rule, hedonistic vs two level, and classical vs. negative utilitarianism, see here.  For a response to the utility monster and repugnant conclusion, see here.  And for a look at whether to count lives not yet in being, see here.

Also, note that I'm now putting page breaks in the middle of my posts so that you can see more than one on the front page...


I'm going to start off by making a note about something slightly different from the content of this post.  Earlier, I defined a philosophy as a preference ordering on all possible universes; the ordering had to be transitive, reflexive, etc.  Basically, a philosophy is something that compares any two possible universes;  in other words, it tells you which options are the best (if you have complete information, that is).  Perhaps for you a philosophy is something different.  Maybe it's something that compares some situations but doesn't say anything about other comparisons.  Maybe it's a binary function that calls all actions either morally permissable or impermissable.  Maybe it is a framwork to look at actions that doesn't necssarily tell you which are best, but instead some other difficult to define properties of them.  Probabily it's a mechanism to justify your current way of life.  Anyway, if you don't think a philosophy should be a preference ordering on possible universes, there's probably very little I can do to convince you, just as if you think faith is more important than evidence or that gut instincts are more important that statistics in baseball there's probably little I can do to convince you.  But from now on I am using that definition, and will look critically upon philosophies that fail to create a preference ordering.

Act and Omission

Anyway, there is a large debate in philosophy about whether taking an action should be treated asymetrically from failing to take an action--the act/omission distinction.  There are many phrasings of the problem, but here is one of the more famous ones: the trolley problem.  The trolley problem is a thought experiment in which you, the actor, are near some trolley tracks.  The tracks split, and past the split there are currently five people tied to one of the tracks, and three tied to the other.  You're standing next to the lever which controls which path the trolley takes; in the first version of the problem the lever is currently such that three people will do, and in the second version of the problem the lever is currently such that the trolley will run over and kill the five people.  A trolley is coming.  You have time to pull the lever, if you want, but not to untie any of the people.  In the first version it's pretty clear you don't pull the lever--not only are you causing the death of five people, but you're only saving three by doing it.  But how about the seond version?  Do you pull the lever and switch the trolley, kiling three other people, or not do anything and let the first five people die?  That is, do you act, or not?  And should morality treat the omission of action, which results in two extra deaths, the same way it would the action of killing two people?  In other words, are these two scenarios the same?  Does it matter which way the lever is currently pointing?

No, it doesn't matter which way the lever is currently pointing.  If you let five people die they will die whether or not it involved pulling a lever and society won't lose any less from their absence than it would if you had pulled a lever to kill them.  To put it bluntly, their corpses won't be terribly consoled by your lack of action in their death.  Death is death.

Now, of course, in everyday life there are lots of reasons that omission of a good action is sometimes not as bad as making a (different) bad one.  If instead of a lever it were some complicated chemical plant explosion with five people in one room and three in another, you might not know enough to know exactly how to save the five people, and might risk accidentally killing all eight if you try to save the five.  And if the cost to you of acting were big, instead of being pulling a lever, that would have to be factored in, and would rationally discourage action.

But it's important not to double count intuitions: these practical considerations are already built in to utilitarianism, and it would only advocate action when, taking in to account all costs of doing the action, it were worth it.  But fundamentally the five people live and die and their lives are no less worthy if they require a lever to be pulled.

It's worth noting, by the way, that the distinction between act and omission is totally undefinable.  In this case there's a general consensus about which is which, but in the end all you have are a list of options--the choices you can make with your life.  Some will involve more physical movement than others, some will involve changing your life more than others, but in the end all you have are your choices.

Creation and Destruction

Another distinction often made is the one between bringing things (or people) into being, and removing them from being--the creation/destruction distinction.  For instance, is murdering someone the same thing as declining to have a kid?

Well, that's obviously a bad example.  Murdering someone kills someone who already had friends, and family, and who society had invested lots of money into educating and training for a job; declining to have a kid is described by none of those things.

Ok, then how about this: is buying someone a new car fair repayment for destroying their old one?

Well, yeah, it pretty clearly is, other than the practical difference that breaking a car doesn't reap the payoff to invert the cost to the actor associated with buying or breaking the new one.  There are also of course feelings involved--people might feel less safe from knowing someone breaks cars--but I think most people  would see this as a fair action on the behalf of the actor.

Ok, well, how about this, then.  Say there are two planets, A and B.  Right now A is barren and B is a thriving metropolis.  They will never communicate.  You can push a button and destroy planet B in order to instantaneously make planet A an even more advanced and happy civilization, trading creation for destruction.  Do you do it?

Well, as you've probably caught on, this is really just a trolley problem in disguise--and that often the creation and destruction issue is really just an act omission one.

There are lots and lots of variants on this--not just act/omission and creation/destruction but also bringing into life versus murdering, giving versus taking, etc.  But in the end they all really boil down to the same thing: if you are walking by a pond and see someone drowning, and know that you could probably save them if you tried, and that no one else will, is failing to act the same thing as murdering someone?  Is failing to do something nice for someone--for instance giving them a good they value, if you want to also make it a creation/destruction issue--the same thing as doing something bad to them or destroying something of theirs?  (This is all assuming they are comparable activities and that there isn't significant social stigma associated with destruction not associated with creation.)

A lot of people feel like there is a difference.  They feel like as long as you're not a dick to people, you don't really have to go out of your way to them--though almost everyone would join in celebrating someone who did (as long as that person didn't make them feel too insecure).  But here you see the issue--everyone agrees that going out of your way to do good things is the right thing to do, they just don't feel like they have to do it.

And that's what the act omission distinction is really about.  Really, for many people it boils down to the question, to paraphrase Parfit: what do we owe to each other?  Do we owe to each other our attempts to save others' lives, or just our promises not to actively kill each other?

But I would argue that that's now how we should think about it--certainly not how a utilitarian should.  Because what matters, in the end, is not what we owe each other.  The drowning man in the pond is going to drown,  if you don't act, whether or not you owe him anything.  You can think all you want about how much karma you've accumulated, but in the end the drowning man will live if you choose to act and fail if you fail to do so--and no definitions of debt or karma or obligation will change that.

And either way you've made a decision.  If you don't save a drowning man you've chosen not to; you've chosen to value the dryness of your clothes over their life.  You can use as passive language as you want to describe it but in the end you are the choices you make and pretending you didn't really choose to act as you did does nothing to change that.


  1. From the perspective of a society, though, non-actors are less harmful than actors after the fact. Someone who doesn't jump into a lake to save someone, disregarding the danger of doing so, is less of a hazard than someone who shoots a person about to push them into a lake. Ultimately, a lot of examples are going to incorporate things society tries to do to aid good decision-making, and prevent people who tend to decrease utility from doing so or punish people who tend to decrease utility. That's the idea, anyway. Being in a world of irrational actors changes how society judges your actions, which changes how you, as a rational or irrational actor, see the act of letting someone die vs killing them. Killing someone not only has a bigger impact on you than killing someone else, making a Monkeysphere-utilitarian more likely to care about killing someone else than letting someone else die. I know you're arguing against Monkeysphere-utilitarianism (which is what I will call the philosophy that personal utility, and that of your friends and family, is more important) but we are, at best, Monkeysphere-utilitarians by instinct. In a world full of Monkeysphere-utilitarians, too, the action of an outsider is a threat to you, and suddenly this person is important and dangerous, whereas the inaction of an outsider is ... meh, that person's not really important, you can't protect your family or friends from lakes and ambivalence. They don't expect more than ambivalence, either, because they're Monkeysphere-utilitarians, too, and they expect most people are more or less like them.

    In a world of actors with these priorities, some of the utilities of inaction break down.

    1. It's certainly true that society isn't utilitarian and that society does make an act/omission distinction, and so e.g. killing someone comes with a prison sentence that failing to save them doesn't; that's another practicle consideration to take into account. I'm arguing from a normative perspective and arguing that there is no inherent difference between act and omission other than the practical ones that are built into society--that is, that utilitarianism doesn't have to have some sort of adjustment factor for acts and omissions because it's already built in to peoples' utilities.

    2. "From the perspective of a society, though, non-actors are less harmful than actors after the fact. "

      I think I disagree with this. I bet there are some really good arguments to be made that there is much more 'not-good' done in the world than 'evil'.

      If one really tried, perhaps they could kill 5,10, maybe 20 people? Meanwhile, an altruist trying hard seems like he or she could save hundreds if not thousands. So on the net, I'd prefer a world of one marginal 'very good' person and one marginal 'very bad' person than two marginal 'average' people.

  2. > But here you see the issue--everyone agrees that going out of your way to do good things is the right thing to do, they just don't feel like they have to do it.

    So, you may be *right*, but I think you're being somewhat disingenuous in simplifying things this far. Just because something is (locally) the right thing to do doesn't mean everyone should be doing it.

    1. Re-reading that line of mine, I wish I worded it differently; it sounds a bit condescending.

      But I still think it's roughly right. It's true that there are lots of things where something is the right thing for one person to do but not another, but I think that most of the important issues most people (or at least I) think about don't fall into that category. You could image that it was correct for me to be vegan, but not for most people, but I don't think that's the case.

      But maybe I'm just seeing problems through that lens and not recognizing lots of important ones... Could you give some examples?

    2. I think what I mean is that acting consumes resources and presents opportunity costs where omission does not. Sure, you can say that utilitarianism already counts that. But I think it's important to note in the context of why people make the distinction (there is, of course, also laziness and selfishness).

      As for a concrete example, well, I'm sure you'd agree that much of the population of the third world is well below your level of utility, and locally the right thing to do (in the interest of going out of your way to help others), for you and each other student at your school, might well be to contribute your wealth and labor in whatever way directly benefits them. But if the whole school did this, I'd wager the world as a whole would be worse off in a generation or two. In that sense, I think your utilitarian calculation might be shortsighted -- it needs to include the resources consumed by acting. If you're a skilled surgeon, then helping people out that you see on the street may cause more people to die in the hospital. I'm not saying saving the drowning man is going to be wrong, just that you need to decide whether it's right using the full balance of utility on either side.

      I realize "but if you did this ALL THE TIME it would be bad" is something of a fallacy, but I'm just trying to make the signs on my examples clearer. Again, I'm not arguing that you're wrong in your conclusions.