The Trolley Problem (2)

In The Trolley Problem, I presented Philippa Foot’s dilemma, which I’ll include here for context:

An out-of-control trolley is rushing down its track - and a mad philosopher has tied five people in its path. There is a switch which will lead the trolley down a different track - but there is a single person tied there. Should you flip the switch?

I also presented Judith Jarvis Thomson’s variant in which instead of flicking a switch to divert the trolley onto a track with a single person, you have the option to push a “Fat Man” off a bridge in order to stop the trolley:

As before, a trolley is hurtling towards five people. But this time you are on a bridge standing next to a fat man. The only way you can stop the trolley is to push the fat man in the path of the trolley, killing him to save the other five. Should you proceed?

Moral dilemmas such as these serve as a fulcrum by which we can explore the ethics of a particular situation, and the post received quite vibrant and varied responses in the comments. I shall set aside those responses that treated the dilemma as the Kobayashi Maru and provided lyrical responses that dodged the issues at the heart of the question (although these were certainly entertaining!)

Foster Nichols provided a decisive analysis:

Assuming that these are the only choices I have (which is, as always seems to be true with these 'dilemmas', entirely unrealistic), that I know nothing about the people, and that I am acting in the heat of the moment, I might flip the switch, but would not push the fat man.

Why? Using the switch is much more impersonal, making the choice seem more quantitative – do I lose five or lose one? Pushing the man is more direct – the question becomes: "Are the lives of the five people more valuable than the life of the fat man?" There's a chance that I would make a decision in the first case, but only if my time was limited.

Foster’s response is the most common in respect of this dilemma; people generally feel it is permissible to flip the switch (whether or not they feel they could do so), and many think it is the best choice in this situation. In the case of the Fat Man, most people feel great resistance to using him to stop the trolley (even if they had no difficulty sacrificing the one in the first case).

Tom Camfield gives an alternative view on the problem:

To flip the switch and not push the man seems analogous to eating beef but being unwilling to slaughter a cow, which seems plain mad to me, though is obviously the case for many people.

I think I'd be happier pushing the man and feeling fully responsible than to flip a switch. Flipping the switch seems too indirect, you could make out it didn't happen, avoid responsibility. I want to be directly responsible for my actions. I think that concept is analogous to the idea that if you're going to kill someone you should do it face to face not shoot them in the back!

Tom’s perspective focuses on a need to be directly responsible for our actions – which leads him to conclude he would be happier pushing the man than flipping the switch. (He doesn’t say that he would kill the Fat Man, but let us assume this for the sake of argument).

I am sensitive to Tom’s argument here – if one is willing to push the switch, does this not open the door to all kinds of problems in the vein of Milgram’s infamous experiment, whereby one thinks less about the consequences of one’s actions because one is only operating a device, following orders, or otherwise has diminished responsibility? I also appreciate the logic that says it may be easier to have something or someone killed than to murder someone yourself, but do we really want to make this process any easier?

The trouble with this approach is the pragmatics of murdering the Fat Man. Darius K picks up on this point:

I do like the subtle twist in the second version. In the first version, where the one person who could die to save the rest was tied to the track as well, I would definitely go for it and flip the switch. In the second version, the fat man has pretty much nothing to do with the scenario in the first place; you are bringing him into the mess by pushing him off the bridge.

Duoae echoes this point:

In the first example you are minimising damage. In the second you are choosing to kill someone who wasn't in danger in the first place.

This indeed is one important distinction between the two scenarios. In the first, the mad philosopher is responsible for placing the six people in danger (one may argue that the single person on the second track is not in immediate danger, but surely the very fact that one person was tied to this track is part of the mad philosopher’s plan). You have the option to save one or five people, by letting the converse number die. In this scenario, a simple calculation seems to allow the death of one to save the five.

But in the case of the Fat Man, he is just an innocent bystander! (Unless, of course, the mad philosopher placed him on the bridge… but even then, he is not in any danger except by your own murderous impulses). As Darius says, you are responsible for his involvement if you decide to push him off.

This reflects what is known as the principle of double effect (attributed to Thomas Aquinas) – which states that it is permissible to act in a way that has an unintended negative side effect, but not permissible to cause harm intentionally. In the regular Trolley Dilemma one does not intend harm to anyone – harming the one is just a side effect of saving the five. In the Fat Man variant, one must choose to harm him in order to save the five, which Aquinas would not have approved of.

Judith Jarvis Thomson, who proposed the Fat Man variant, argues that in the original problem we deflect the harm, whereas in the second case we cause harm. She argues in the first case that nobody has more right than anyone else not to be run over, but the Fat Man does have a right not to be pushed in front of the trolley.

From a legal perspective, Thomson is absolutely correct. If we imagine the court case that results in either situation, it seems likely (barring courtroom gymnastics!) that you would be acquitted of murdering the one to save the five in the case of flipping the switch. Conversely, it seems likely that a charge of second degree murder would result in your conviction if you decided to kill the Fat Man to stop the trolley. The families of the five might consider you a hero, but society would not.

Utilitarians (a particular ethical system which is wholly outcome-focussed) deny that it can make a difference whether you murder the Fat Man or flip the switch, because their ethics are based solely on assessing the results of actions. I will argue against the desirability of this form of ethics at some point in the future, but at the very least it can be objected that this logic is what allows our governments to commit atrocities on our behalf.

Tom expands his position (which allows him to murder the Fat Man) as follows:

The utilitarian decision is "1 death is better than 5", the Kantian one is "I ought not kill", but both those systems take away the responsibility from the individual ie "it's not my fault they died, the system of morals told me how to act!" I think my position is closer to virtue ethics and existentialism, and implies that the freedom of the individual to act and the necessity of personal responsibility is paramount. That doesn't make it better, just bases the decision on different principles.

I find it interesting that Tom berates both the utilitarian (outcome-focussed) and deontological (rule/rights-focussed) approaches as taking away individual responsibility. For myself, I see the problem with the utilitarian approach – it permits the individual to murder the Fat Man – but not with the Kantian position. The Kantian interpretation is not simply “I ought not kill” (although Kant did believe this personally), but rather that to murder the Fat Man is to use him solely as a means – to deny him mutual respect. I contend that this denial of mutual respect invalidates Tom’s justification for murdering the Fat Man – the freedom of the individual to act and the necessity of personal responsibility cannot allow us the capacity to overrule these rights in other people, except from an egoist’s position which is dubious in ethical terms.

As Foster notes:

It really shouldn't be my decision at all, though. The fat man (or single person) should decide for himself.

As for acting in respect of a system of rules, I may agree that to not take action because of a strict deontological system of ethics (e.g. “I will never kill”) attempts to excuse the person of their personal responsibility, but pragmatically it does not alter that responsibility in the context of society. If one chooses to honour a system of rules, that is your choice, and you remain responsible for your choices. I believe deontology is not incompatible with existentialism as some degree of choice, and some manner of personal interpretation, is always required in moral matters.

Duncan reflects upon the dilemma in a vastly more human manner:

Realistically, faced with a life or death situation, my logical side kicks in. If I knew none of the victims, then the one over the many. However, if the one is someone important to me (say, my wife) then emotion wins and she lives. If there is positive emotional connection on both sides, they get to duel it out (probably into inaction).

Here we can see clearly what I believe the Trolley Problem demonstrates in respect of outcome-focussed ethical systems, namely that when all else fails we can use an outcome-focussed system as a “tie breaker”: if I know none of the victims, it is easy to say five survivors outweighs one survivor. But when we know the people involved, we save the people who are important to us.

Duncan doesn’t expressly say that he would not murder the Fat Man, but he does comment that he might try to persuade the Fat Man to sacrifice himself.

He is not the only one to express concern that their actual behaviour in such a situation would be inaction. Foster again:

Either way, the dilemma really boils down to the question about relative value, which I don't have an answer for. (Are five always worth more than one? Is being 'fat' bad, or just convenient for stopping a trolley?) That, along with being very unwilling to kill someone, would probably cause me to do nothing in both situations. I realize that people die no matter what I do, but I think there's a significant difference between letting people die and choosing to kill someone else.

We have already seen at least one such “significant difference” between letting people die and choosing to kill, namely the legal consequences. The fact that utilitarian ethical systems do not see this difference suggests to me a flaw in ethical systems that judge moral matters solely from outcomes.

Before we finish, we should look at how Duoae recasts the dilemma into an interesting but very different form:

1) You are tied to the track. The cart is headed towards you with five people onboard. You can reach a switch that will divert them off a cliff and leave you alive. If they hit you they will slow down and stop safely.

2) You're still tied to the track but this time there's a fat man on the other side.

Here we are exploring whether you are willing to sacrifice your own life to save five, or murder another rather than save oneself! These are very different questions, and I leave it as an exercise for curious readers to explore for themselves.

Finally, Marc takes a wry look at the whole issue and suggests:

Don't throw the switch, don't push the fat man. Let the trolley hit the five people in either case - the magnitude of the tragedy will bring the spectre of horrific runaway trolley accidents into the public eye, forcing the trolley operators to implement and enforce stricter safety regulations, preventing untold trolley-related deaths in the future.

My thanks to everyone for their fascinating and thought provoking responses to this dilemma, and I look forward to exploring more such problems in the future!

Comments

[Corvus]

My problem with this scenarios is that they toggle my "flippant" switch. Probably because I've had too many people ask them without any intention of in depth analysis. This is a good follow up and made me realize something:

If we replaced the switch with a vote and the trolley with foreign policy... it'd be a very similar conversation, wouldn't it?

[Jack Monahan]

All I can say after thinking through these scenarios is that I'd like to avoid ever getting on or being near a trolley--clearly a deathtrap for someone! :)
That and thank you for the link on the Kobayashi Maru. I watched enough TNG, but not enough of the original series to catch the reference personally, but now it's just another in the bag of geekdom lore.

[John Peacock]

Don't intervene at all, but record the whole sorry event on your phone or camcorder and auction it to the highest-bidding news-gathering organisation.

Then sue the tram company for emotional distress at having to witness such a horrible event.

Modern morality: it's not very moral, but it's terribly modern.

[Chris]

"Modern morality: it's not very moral, but it's terribly modern."

John, I couldn't agree more! :)

"If we replaced the switch with a vote and the trolley with foreign policy... it'd be a very similar conversation, wouldn't it?"

Well I'm trying to avoid the political on my blog at the moment, but the parallel you raise is apposite, Corvus!

Oh, and Jack, "the bag of geekdom lore" is vast and abyssal. Be careful when you open it that it doesn't swallow you whole. ;)

[zenBen]

"We have already seen at least one such “significant difference” between letting people die and choosing to kill, namely the legal consequences. The fact that utilitarian ethical systems do not see this difference suggests to me a flaw in ethical systems that judge moral matters solely from outcomes."

This seems a shallow appraisal of the utilitarian position that smacks slightly of your bias against it. Of course the utilitarian will consider the legal ramifications, if they are aware of them, because they are part of their own ethical considerations and must account for themselves. So they would also have to consider socially-based feedback, such as gratitude or revenge of participants and their acquaintances; the attitude of the media, if there is any in that society; possible political ramifications (perhaps depending on the race of participants - is the fat man black?).

Is it meant to be ironic that your trolley picture has tyres (so does not run on a track)?

[Chris]

zenBen: It did make me chuckle that the best picture of a trolley I could find was one that couldn't run on tracks. :)

We will address my opinion of utilitarianism in due time. However, in brief, Rawls demonstrates relatively conclusively that it will not produce a viable formal system, Arendt is justly critical of its premise, and its central conceit - that we should "maximise utility" or "act for the greater good" - is a dubious justification, capable of producing serious atrocities when left in the hands of politicians.

I am certain that there are forms of utilitarianism that can be 'rescued' and made viable, as its positive motivations are laudable, but this is a moral philosophy in serious crisis right now.

If you know of a twentieth or twenty first century moral philosopher with a solid take on utilitarianism, by all means point me to their work - I could use to see a reasonable counter argument right now.

People are free to choose to be utilitarians just as they are free to choose to be egoists, and I won't argue otherwise. But I have grave concerns about the effects utilitarianism has had in casting the shape of the modern world.

I look forward to discussing this in more detail in the near future! :)