Thought experiments

[qatarperegrine]

I've been having fun debating ethical thought experiments with my friends lately, and I figured I'd share the love right here on LJ.

Here's the first set of scenarios, often called the Runaway Trolley Car experiment. I'll steal the wording given in a BBC article on ethics.

1. A runaway trolley car is hurtling down a track. In its path are five people who will definitely be killed unless you, a bystander, flip a switch which will divert it on to another track, where it will kill one person. Should you flip the switch?


2. The runaway trolley car is hurtling down a track where it will kill five people. You are standing on a bridge above the track and, aware of the imminent disaster, you decide to jump on the track to block the trolley car. Although you will die, the five people will be saved.
   
   Just before your leap, you realise that you are too light to stop the trolley. Next to you, a fat man is standing on the very edge of the bridge. He would certainly block the trolley, although he would undoubtedly die from the impact. A small nudge and he would fall right onto the track below. No one would ever know. Should you push him?

So here's the major question: was your answer the same for both scenarios? And if not, why not? Can you rationally justify why the scenarios might call for different responses, even though the results (one death or five) are the same in each case?

Comments

[(Anonymous)]

a) yes
b) no

No. Because I am selfish and want to live forever.

[rmitz.livejournal.com]

Oops, I wasn't logged in.

[qatar.livejournal.com]

I think you missed a paragraph -- it's the guy next to you who bites it, not you.

[rmitz.livejournal.com]

Oh, I did miss that. That makes it harder to answer. :) Assuming he is a stranger, I'd like to make the decision in the same way. I don't know if I actually could though.

[qatar.livejournal.com]

By "in the same way," I assume you mean "in the same way as scenario 1" and not "in the same way I did when I said I didn't want to die." ;-)

[seetarkrun.livejournal.com]

a. yes, though I'd be tormented about killing that one person

b. I guess the difference is that in a, everyone died the same way - you're only changing the number of equally painful deaths. But pushing someone off SEEMS like a worse death. Plus, there's the obvious pushing a button vs. pushing a person distinction, the latter of which seems much more insidious.

Plus in b, you're discriminating against fat people. Sort of.

[qatar.livejournal.com]

I wonder why it seems like a worse death. You mean because he has to endure falling off a bridge as well?

To me that fat man's death seems worse because it seems more unfair than the death of the lone person on the track in scenario 1. He isn't on a railroad track; he's just a random bystander. His life isn't inherently in danger until WE come along and intervene. (Justin says: are you saying that people who walk on railroad tracks deserve what's coming to them? No, obviously.) I don't think that this is a rational reaction to the situation, but it's one of my gut reactions.

Other than that, yes, I think there's a huge difference, emotionally, between pushing a lever and pushing a person. (Research corroborates the idea that people respond much more emotionally to the second scenario than the first.)

[kit-ping.livejournal.com]

Justin says: are you saying that people who walk on railroad tracks deserve what's coming to them?

Someone who is walking on a trolley track has acknowledged and accepted that his future might include being hit by a trolley. The reason it seems less unfair to take out the man on the tracks is that dying that was was a risk he himself had decided to take. By staying on the tracks he has, in an odd way, given a trolley permission to run over him. The man on the bridge hasn't.

[qatar.livejournal.com]

Yes, that's what I thought too. Justin's point: what if the man on the tracks had KNOWN that the switch was thrown the other way? Then he is not taking a risk.

And isn't a man standing on a footbridge incurring the risk of being pushed off? :-)

(I need to make myself a devil's advocate userpic.....)

[qatar.livejournal.com]

There, that'll do for now. Took me forever to upload the dratted thing -- Photoshop loves to bloat images above LJ's 40K limit!

[lasa.livejournal.com]

That's Screwtape, isn't it?

[qatar.livejournal.com]

Wow, you're good!

Yes, the image is stolen from the most recent edition of Screwtape.

I was thinking of using a picture of Sweet, the demon from the musical episode of Buffy, but I couldn't find a good screen capture. :-)

[kit-ping.livejournal.com]

If he knows enough about trains to know that the switch was thrown the other way, he knows enough to be aware that emergencies do arise that might cause a train to be shunted onto his track. Plus, lacking camera feed or line of sight, he doesn't know the switch is thrown the other way; he just knows it was when he checked. Change happens. (And if he did have a camera feed or line-of-sight, he'd have been able to get out of the way! :)

A friend of mine has a userpic that's a picture of a pair of beat-up jeans with an angry face spray-painted on them and text that reads, "Uh-oh! Gywn is wearing the CRANKY PANTS again!" I should steal it for debating; the longer I go the more curmudgeonly I end up sounding. :)

[foobart.livejournal.com]

Ok, curmudgeon:

Say the bridge is actually a trolley bridge, and the fat man is standing on a different (at the present instant, unused) trolley track, meaning he is taking the same risk as the single guy in the first example. Does that change your answer?

If not, I think you're inconsistent. If so, I think you're absurd. :)

[kit-ping.livejournal.com]

*reads*

*rereads*

*looks at you*

*goes to get something from the oven*

Here, kid. Have a cookie and be quiet. ;)

[foobart.livejournal.com]

Yum!

[seetarkrun.livejournal.com]

It seems like you're forcing the man on the footbridge to make a sacrifice, where the person standing on the tracks, while making a sacrifice as well, isn't making much more of one than the other five would have. That made more sense in my head, I think.

When I first saw this, I thought it was going to be the one person you love vs. 10 strangers dilemma. I'm very glad it wasn't, because that's not a decision I think I could make...

[chrisamaphone]

yeah, they showed us this thought experiment in my cogpsych class 3 semesters ago. though we had a third scenario, too:

You're in a hospital and there are five people who desparately need replacement organs, all different. They'll likely die that day if they don't get them. There's a guy who's visiting the doctor for his daily checkup, and he's perfectly healthy. It turns out that he could provide all the organs necessary for the five terminally ill people. Is it morally ok for his doctor to cut him into little bits for the five people?

[qatar.livejournal.com]

That was going to be my next thought experiment! :-) It's usually paired with a similar scenario, like the runaway trolley scenarios are:

You're in a hospital and six mortally injured people come in for treatment. One is much more seriously injured than the rest. If you save that person, the other five will die for lack of treatment. If you treat the other five, then you won't have time to save the one. Who do you treat?

As in the trolley experiment, most of us would choose to treat the five and let the one die -- but most of us would NOT choose to kill the perfectly healthy person to harvest organs for the other five. In both thought experiments, though the net result is the same (five deaths vs. one) each scenario feels very different.

In this one, though, the act/omission distinction is even more clear. Most of us find a big difference between allowing a mortally injured person to die and actually killing a healthy person.

[qatar.livejournal.com]

To clarify: most of us SAY we would choose to treat the five and let the one die. I've a suspicion that for many people it would probably depend greatly on who the five were and who the one was. (Would it matter if the five were career criminals? Would it matter if the one was your daughter?)

[chrisamaphone]

oh, sorry -- didn't mean to ruin your next thought experiment!

interesting, though; i hadn't heard it paired with that scenario before (for us it was alongside the trolley scenarios).

[chrisamaphone]

(I guess I should remove the "likely", though, for it to be analogous to the other scenarios.)

i think a lot of this has to do with how sort of "close up" to your deed you have to be to make the change. it's still pretty baffling to me, though. i dwelled on this for days when i first heard it.

i guess really i'm not okay with _any_ of the three situations. i dislike the idea of sacrificing someone's life for others when the person hasn't expressed a desire to make that sacrifice. even throwing the switch in the first scenario seems horribly unfair.

it's possible that we make an _assumption_ that the one guy tied to the rail would be willing to make that sacrifice, whereas for the others it's more obvious that they'd object.

[(Anonymous)]

1. No
2. No

If I just happen to watch it happen - it's fate. If I do something to make it happen or happen slightly in a different way - I am a killer. Argh.

Qatar Cat

[qatar.livejournal.com]

"If I just happen to watch it happen - it's fate."

So we have no ethical responsibility to STOP bad things from happening? Even if you can intervene to stop someone's death, it's OK not to?

[(Anonymous)]

I think I'd rather not do anything than do something that will definitely result in loss of life. I wouldn't be able to sleep at night knowing that I purposedly killed a man. Even if I saved 5 instead. Those five would not be killed by me, but that one guy would. I couldn't live with that.

Qatar Cat

[kit-ping.livejournal.com]

Sorry, my brain got stuck on the statistical improbability of being the too-small person on the bridge. ;)

There are times when I wish that I hadn't so thoroughly chucked my Mentor books, such as whenever things like this come up. It's possible that, after tea, I'll be able to remember details about the study of differneces in the way men and women tend to analyze ethical quandaries (aka social web vs abstraction, aka why the problems-in-a-vacuum we had to discuss in Mentor III always drove me bugf*** insane) and why women, especially women denied schooling (as most were when these compasses were being developed and propagated), always seemed to be unable to get to the highest category of ethical thought.

It has to do with the "does it matter if..?"question that someone else brought up down below. My "typically female" response: YES, it bloody well does. It always does. The situation is changed if the five people have guns trained on the one and have been taking shots at him in such a way that his terror and pain will be drawn out for as long as possible. The situation is changed if the fat man has an EMT's pack over his shoulder and is heading for an injured child just up the bridge. The situation is changed if the lone person is the landholder who bribed public officials to take your family's farm out from under you, causing your youngest daughter to die because you could no longer afford to treat her rare disease.

This isn't "extra". It's information vital to making a complete and informed decision that has been deliberately withheld in the original scenario. Anyone who who says a full conclusion can be reached without full knowledge is being intellectually dishonest. [Relevant but highly combative concluding sentence removed for the common good.]

Yeah. My Mentor III class was ever-so-much fun. I'm not sure who was happier when I was done, me or the professor.

[qatar.livejournal.com]

You're thinking of Carol Gilligan's criticism of Kohlberg's stages of moral development and, I think, propounding an "ethics of care" over an "ethics of justice." :-)

I can't argue that you need to have all the information vital to making a complete and informed decision. But I think the point of these thought experiments is to figure out what the myriad responsibilities and obligations we feel ARE, by constructing scenarios that differ in only one aspect (e.g. do you kill the person by commission or omission) and reflecting on whether that changes our reaction. If we only thought about ethics in context -- if we always knew who the one person was, and who the five people are -- then we'd never know if we made that decision because we discriminate between acts of commission and acts of omission, or just 'cause we liked one group of people better than the other.

I do accept the point of view that ethics can't be considered out of context, though.

(Is it then fair to set up a scenario in which you can't possibly know the identities of the six? Because there surely are situations when an ethical decision has to be made without knowing all of those non-extra pieces of information. As, say, when a government's actions may make a difference between one Israeli dying or five Palestinians. Or do you think the decision can't be made without a full dossier on each individual involved?)

On a more humorous note, I think you would enjoy this thought experiment, which combines every thought experiment I've ever heard and then some. ;-)

[kit-ping.livejournal.com]

First off, I don't know what was sillier, the brain inna jar setup itself or how many commenters didn't get the joke. :)

You're thinking of Carol Gilligan's criticism of Kohlberg's stages of moral development and, I think, propounding an "ethics of care" over an "ethics of justice." :-)

Yep, that's the one. We read an excerpt from it, after which I checked out the whole thing from the library and settled in to seriously annoying my classmates. ;) (It helps to know that, because of the time slot, my mentor III class was entirely school of business students plus three us us from the conservatory. Let's just say that people who gravitate towards business management and people who gravitate towards the arts have differing views of the world. I'm sure there are socially conscious businesspeople given to reflection and thought out there somewhere- they just weren't in that room.)

I agree that these are useful as a starting point. But that's all they are. There's an annoying tendency, especially in, for lack of a better word, layman's discussions, to presume that the reaction changes based on the one variable are somehow more pure (eh, maybe) and therefore more valid (er, no) than changes based on circumstance. That somehow taking into account the details must mean reacting solely on emotion. Pah, sez I. (You'll note I that deliberately outlined changes which weren't based on if you liked them or not, but on their actions: a great wrong currently being committed, a great good about to happen, a past injustice unpaid for. I agree, using whether or not someone irritates me as a reason to let them be squooshed by a train is somewhat suspect. :)

Is it then fair to set up a scenario in which you can't possibly know the identities of the six?

I just feel that when you try to strip it down that far it gets kind of ridiculous. Take the government action: Even if you-the-government just know that there's one Israeli and five Palestinians, that gives you a bunch of information. You know who's allied with you, who's most recently ticked you off, who's more likely to cause substantial damage in response, who has other allies you don't want to annoy... and you'll use that to make a decision.

True anonymity is so highly unlikely that it's almost self-defeating. I think that's why questions such as this are posed in a real-life context. Make it a possibility and people will see the need for thought and discussion; bring it down to "one person dead or five people dead. Choose." and the question is more likely to be "why should I?" than "What should I do?"

Did you ever see The West Wing? For some reason, I pictured the exchange being played out between Leo McGarry and President Bartlett:

L: Okay, Mr. President, we need your order on whether to target one person or five. Now.
B: Why?
L: Because.
B: Just "because"?
L: Yes, sir.
B. No.

do you think the decision can't be made without a full dossier on each individual involved?

You cannot make a fully informed, fully ethical decision without all the information, no. It's a highly imperfect world, though, so you take what you have, use it, and hope for the best. But until you fully understand the situation you'll never be able to definitively say it was the right choice.

[nicodemusrat.livejournal.com]

While I agree with my wife's upper comment on acknowledgement of risk by the people walking along the trolley track, I think there's another factor. In changing the track switch, one can be almost entirely certain that the trolley will be diverted and may kill one person. In pushing a person off a bridge, especially since the weight difference between me and the larger person is evidently critical, there's a great deal more uncertainty in whether this will be capable of stopping the trolley (or whether your aim will be good enough to land him square on the tracks, etc.).

So the second situation involves a greater chance of failure, in which you'd be adding an unnecessary death on top of the five people in the trolley's path.

[qatar.livejournal.com]

The thought experiment asks you to imagine certainty. You are SURE that his death would stop the train.

Whether we can really convince our brains to believe that is another matter, I suppose.

[nicodemusrat.livejournal.com]

Exactly. I got caught on the question of why I was perceiving these two scenarios differently. My brain was trying to disbelieve the assumptions, which created an apparent differentiation.

[foobart.livejournal.com]

I think this is a cop-out.

I don't think you can escape the dilemma by calling the outcome uncertain unless you believe the uncertainty is inherent in and core to the basic idea. In other words, unless you believe that it's impossible to construct a scenario to ask the same basic question without having the same amount of uncertainty, the issue has been dodged, not addressed.

For instance, let's say the scenario was thus:

Dr. Evil has placed 5 people in a pit, which is slowly being filled with water so that the sharks with frickin' laser beams can be released. Dr. Evil's incompetent henchmen have left a pulley in the ceiling above the pit, which is threaded with a lightweight but strong cable. You have one end of the cable, and the other end is in the pit.

Each of the people in the pit weighs 130 pounds. You weigh 110 pounds. As the pit fills, someone you've never met wanders into Dr. Evil's lair. It's Jared, who has recently switched from an all-subway diet to a diet consisting entirely of deep-fried snickers bars and sugar-coated lard. He now spends his days in channel surfing looking for tv ads in which he appears. As a result of his lifestyle change, he tips the scales at 700 lbs.

You have a choice: do you kill/incapacitate Jared and use his body to pull the 5 people to safety, or not?

[nicodemusrat.livejournal.com]

I don't think it's a cop-out because the original post's question revolves why might we perceive these scenarios differently. I perceive them differently because one seems inherently or intuitively less reliable, even if the scenario asks us to believe otherwise. So the answer is to do with perception.

Given your scenario with attendant assumptions (I know Jared's heavy enough, I know the sharks will be released and kill the people in the tank --why not just shoot him, dad?--, I don't have any emotional favoritism between the people, I wouldn't face legal consequences, etc.), THEN I can say that I would sacrifice Jared. My answers for the two scenarios in the original post would also match.

I just thought that the question of why the scenarios appear different was more interesting.

[(Anonymous)]

I don't think it's really about the saving of 5 people at all. If you weren't there, they'd be killed. In the first scenario, it's more about how you'd feel about being directly responsible for the death of someone by your actions rather than by your inaction.

Another possible reason for different answers to the two scenarios could be the difference between killing someone by shooting them from a distance, or doing it with your bare hands.

By pulling the lever, you're once removed from the act of killing as opposed to pushing the man off the bridge.

Mise

[thebenedictine.livejournal.com]

My emotions make me more likely to shove the person in the way than to switch the tracks.

I wonder why I think that way. I will have to investigate.

[qatar.livejournal.com]

Well, I have to admit nobody else I've asked has come up with that combination of answers! :-D

[thebenedictine.livejournal.com]

the first option equates choosing whether someone lives or dies with a maximization problem, like an equation to be solved by cold robot calculations

the second option makes it my decision, and one that i have to make as much on impulse as on philosophical and moral justifications

i would feel dishonest switching the trains like i wasn't 'really' killing someone

pushing someone directly into a train to save lives, and then like in U-571 sending some young submariner to drown while fixing to sub to save the rest of the crew's lives, you have to take responsibility for that. that's a part of your character.

in the first option you can say 'well anyone would have done it'. that is why i am ambivalent about it, it doesn't say anything about who i actually am, just how i would appear to other people.

i don't know why i'm going at it in this manner, it might be because i've heard the first problem in the context of 'some ethics professor once said to me' or in a book. it's like the 'quintessential problem' or something, and i guess talking about it feels ingenuine to me, dishonest, like for all the talking we can do about it to figure out why we might feel a certain way, why we might want to appear to others in a certain way, it is too removed from emotion to be an accurate depiction of behavior.

wow it's late and i'm running on

[qatar.livejournal.com]

"it is too removed from emotion to be an accurate depiction of behavior."

I'd be curious to find out how people's ACTUAL behavior in ethical dilemmas compares to how people talking about thought experiments SAY they'd behave.

Probably CMU's Institutional Review Board would not approve research that involved making participants shove people in front of runaway trolleys, though!

[lasa.livejournal.com]

I suspect it has a lot to do with how close you feel to the death whether it's justified or not. Pretty obvious, but I'll explain anyway.

Silly example. I went to a 4-H auction on Friday. Here are kids who have raised lambs for six months, who clearly cared about them. I saw them in the barn before the auction hugging them, tenderly putting ribbons around their necks and brushing them - before they brought them to the auction block to be sold by the pound to supermarkets. I got all emotional thinking about those kids and those lambs and how hard that must be.

The next day, I was in Costco, looking at a package of lambchops in shrink wrap thinking, "Those look good..."

Clearly, my consumerism, if I had bought those lambchops, would be a direct contributor to the death of lambs. But see, I didn't see those lambs or the people who raised them. It was meat in shrink wrap.

So - my hypothesis? The further removed you are from the specifics of the death, the easier it is to accept the death.

[qatar.livejournal.com]

That makes sense. So it's easier to live with yourself if you pressed a lever than if you manhandled someone to his death?

Or do you mean that it's not even that it's easier to live with yourself, but that it takes mental gymnastics to even convince yourself that the situations are analagous -- like it takes mental gymnastics to look at a lambchop and think of sheep?

[lasa.livejournal.com]

I think it takes mental gymnastics to convince yourself that the situations are analogous.

It's a bit of self-deception, to make yourself feel better, perhaps? I find the comments above interesting - that the self-deception makes the switching of the tracks scenerio less ethical than the pushing off the bridge scenerio. I can see the point.

I'm imagining a military board room where the strategists are deciding to send in a battillion to get creamed in order to distract the enemy away from their true objective, and thus, end the war more quickly. It'd be a lot easier to make that decision in a board room than in the field surrounded by the doomed soldiers.

[y-pestis.livejournal.com]

Absolutely. That's why my second-level manager made her underling lay people off instead of doing it herself. It's easier when you're removed from it.

I find myself being very aware of this process in food choices, as you mentioned. It's harder for me to eat something that looks like what it originally was (ie shrimp vs ground beef) and I think it's a combination of two elements: one is that more shrimp die to create a shrimp meal (30 lives instead of one - whether the lives are equal is another interesting question) but sadly enough the bigger thing for me is that it's hard to eat something when I've got such a stark visual reminder in front of me of what it was.

After many years of not being able to eat chicken wings or drumsticks, I eventually realized that's why - chicken strips are processed enough that I can avoid thinking about where they came from, but drumsticks have tendons and ligaments and biological details that are hard to avoid thinking about when you're picking them out of your teeth.

Apologies to the vegetarians reading this, who are probably retching right about now.

[qatar.livejournal.com]

I'm not retching; I'm wondering if you remember the stage of my protovegetarianism in which the only meat I would eat was tuna, sausages and pepperoni.

Part of it is that I really liked tuna, sausages and pepperoni, but in large part I think it was because those foods don't look like cute little fuzzy animals.

[y-pestis.livejournal.com]

Hmm. Good point. Although, for the record, shrimp SO don't look like cute little fuzzy animals either. Cockroaches of the sea, yes. And that's a great analogy to bring up when someone's eating lobster, by the way.

[qatar.livejournal.com]

I suppose "fuzzy" isn't a necessary condition for "cute," though. Remember how cute those clams were when they were spitting at us, that one time Mum got clams?

[y-pestis.livejournal.com]

No, I'd blocked that out. Thanks for bringing it back up, now I'm going to have to go see a therapist...

OK, maybe alleged cruelty to molluscs isn't worth a trip to the therapist... but I'm on Kaiser so it's only $5.

The little spitting clams were very cute... I remember thinking it was an unselfconscious defense mechanism, like the way the hamsters would stand up on their hind legs and try to be still and invisible, but actually wobble quite significantly.