Bryan Caplan  

A Trillion Nazis Versus the Trolley Problem

The Book Mike Huemer Should Wr... EconLog Book Club: For a Ne...
Last week, I defended the usefulness of moral hypotheticals.  Last night, I dismissed trolley problems as "silly."  Fenn, an Econlog reader, is understandably puzzled:

"Silly trolley problems?"

Wasn't it just a coupla days ago you were talking about Nazis and defending this kinda thing as the philosophical equivalent to controlled experiments?

The reason why trolley problems are silly isn't that they are hypotheticals.  They are silly, rather, because they're not designed to have clear answers.  As a result, they fail to perform the key function of hypotheticals: clarifying complex moral questions.  And since literal trolley problems almost never happen, they're of no intrinsic interest either.

Comments and Sharing

COMMENTS (12 to date)
Daniel writes:

I think you're wrong about their not being designed to have clear answers. As a matter of fact, perhaps they don't have clear answers, but I think you're confused about the intentions of those who use them.

Initially, the organ transplant case was meant to make trouble for utilitarianism by showing that, contra utilitarianism, killing one is worse than letting five die.

The trolley problem (in which pushing the fat man is contrasted with flipping the switch) was meant to make trouble for that diagnosis of the transplant case.

Precisely because it's supposed to be clear that pushing the fat man isn't ok, and flipping the switch is ok, even though they both amount to killing one to save five, we're supposed to conclude that it can't always be wrong to kill one to save five. So our explanation of just what was wrong with utilitarianism (i.e., that it failed to recognize the morally significant distinction between killing and letting die) doesn't look quite right anymore.

The literature takes off from there, and the aim behind it is to come up with a nice principle that entails that pushing the fat man is wrong, flipping the switch is OK, and doesn't say counterintuitive things about other cases. People present cases that they take to be clear in order to provide counterexamples to previous attempts to provide such a principle.

So the cases are designed to have clear answers. As a matter of fact, it may be that they often don't, but that's a failure of the participants to debate to carry out their project--not part of what they're aiming at.

Lest I seem too sympathetic to the debate, let me not that I generally lean utilitarian, so I'd be ok with pushing the fat man, and I'm skeptical that our gut reactions to these cases are coherent enough to be captured with a simple, elegant principle.

But I'm not sure what project you think the people in that debate are engaged in if you think the cases they discuss are explicitly designed to be unclear.

Fenn writes:

Thanks for addressing my comment, though I still don't quite get your point. The trolley problem seems to equate to transplants about as well as Nazis having a quantifiable hatred of jews does to any real world situation I can think of (displacing indigenous people, maybe?). Not sure I see either problem having clearer answers, either.

But I'm a complete philo newb, so I'm probably missing something. You do seem to lose me where the rubber meets the road on these things: the endowment effect and your kids for example.

But thanks for the pointers and the food for thought. Gonna give your essay on free will a go.

Chuck writes:

I agree with Daniel.

Tom West writes:

clarifying complex moral questions.

And complex moral questions are supposed to have clear answers?

The reason that they're complex moral questions is that they multiple sides and all sorts of trade-offs. The "right" answer is formed from a weighing the different costs and benefits, and of course, that weighing varies from individual to individual. Meaning there's no single "right" answer.

Finding an example that "clarifies" a complex moral question probably means illuminating a single aspect of the question allowing it to have a "right" answer and ignoring everything else as unimportant.

Clever? Yes? Illuminating? Not really.

There's a reason these questions have been pondered by humanity for generations... They have no simple solution (at least for most of humanity).

Tom writes:

I think Bryan is right and the trolley problem is silly and doesn't have a clear answer. If we extend the problem with more information it merely demonstrates ones preferences and I don't see how my preferences are clearly superior to someone else’s.

What if the five on one track were hideous looking people and the one on the other was extremely attractive? In this case would it be all right to let the five perish in favor of the one. What about if the 5 were terminally ill and the one was a small child? Now, would you save the one or the five? What if the five were AIG employees and the one was home owner in foreclosure? Would you still save the five?

Should additional information make a difference in your decision and wouldn't your decision be merely your preference and not necessarily a moral truth that is clear to everyone?

Also, I like the other example where you push the fat guy over. It would be a fat guy since they are not attractive people and therefore worthless. Why couldn’t it have been the muscle bound gentleman standing next to you? Would you try to push that guy over to stop the train or would you be afraid that you would be the one lying on the track if you tried?

Arare Litus writes:

"they're not designed to have clear answers"

But does not the problem clarify what strong utility weighing really means? That you look at the option with net gain and go with that option - and that this abstract calculus can have real effects, i.e. impinge poorly on a minority? Is that not the point, and does the problem not get it across? The toy model cannot be used more broadly, so in a sense it is useless and does not move you forward, but it does clarify, does it not? [the question also relates to the fact that everything is a choice, even the option of doing nothing].

As for not happening in real life - we allocate resources all the time, which has consequences. How we allocate those resources and make decisions is important.

How about this: "for the cost of keeping a pedofile in prision for one year 10 childrens lives could be saved in a typical poor country, the summed cost of keeping the pedofile in prision for his entire sentence would give those children the basic food, eductation, and opportunity to thrive" (numbers made up, and you can bicker on if this is plausible - the thriving - but the general idea is there). Switch pedofile for "victimless crime prisoner" to think about different implications, or change the spending side, or reduce taxes insead. Etc.

This is "real", in that we do have prisioners, and many of them for no apparent constructive reasons, and the resulting use of resources prevents what could have been. Again though, this is a toy problem somewhat - but does it not clarify some specific aspects of the questions involved? And as the questions are related to what we are currently doing, and our stances on how to make decisions, are they not important - even if extreamly limited and "one off" validity?

Isak writes:

No, the trolley problems really do clarify things. For example, when it comes to the case where one is asked to sacrifice the fat man tends to suggest that most people have a problem with sacrificing "innocent bystanders", people who one could not say "knew the risks".

Snark writes:

The problem with trolley problems is that they become hyper-extensions ad infinitum.

What if the five on one track are all ex-convicts and the one on the other track is a doctor on the brink of discovering a cure for cancer? What if the fat guy had previously saved your child's life?

Trolley problems keep us stuck in the rut of moral relativism, where right or wrong have no meaning.

Christiaan writes:

The only real result of these trolley problems is to show a person's individual moral preferences, but does not clarify in any objective sense what is right or wrong.

I think good moral hypotheticals should aim at doing just that, portray some extreme situation where any action besides the 'right' one is blatantly immoral.

Also the utilitarian perspective on these problems is usually quite fallacious. Any measure of utility can only be subjective, thus without enough information regarding all the people on the track's utility to the guy pulling the lever he can't weigh the value of his possible actions. Which means people tend to base their decision on moral self-respect, ie. which action increases their self esteem the most (this includes his perception of other peoples respect for him) All of this makes any answer rather convoluted and certainly does not clear anything up.

ChrisA writes:

If you believe in evolution then you must accept that human beings evolved moral "senses" or instincts in order to allow us to function within groups. Hence it should (and is) very easy to create hypothetical gotcha's like the trolly problem that, like the vase:face optical illusion, cause confusion for our moral senses, flipping from one position to another. Fun, but trivial. We are like ants, suddenly becoming intelligent and mistaking their service of their queen for something noble, when all it is is a reproduction strategy.

Kurbla writes:

I do not see any problems with Trolley problems. But I dislike word "silly."

Matt writes:

How come no one ever seems to account for failure of ones actions in hypothetical situations. Even in most extreme cases when you account for the possibility of error in utilitarianism, there rarely seems to been anything illuminating from it. For example you have no clue pushing that fat guy will stop the train. You have no way of knowing that they five people on the tracks can't be freed. In which case you can easily kill a person completely unnecessarily. This angers me almost as much as the idea of quantifying happiness is as bogus as ideas come. All utilitarians though still need to watch out and obey me, the utilitarian monster!!! I enjoy dominion over you and easting you too, more then you can even comprehend!!

Comments for this entry have been closed
Return to top