Off Your Trolley: Re-Examining A Famous Ethical Dilemma

In my last Blog we talked about John Harsanyi’s brilliant ‘amnesia' principle, and how we can use it as a short cut to tackling ethical issues.  Perhaps the most famous of all the ethical dilemmas in circulation is the trolley problem:

"A trolley is running out of control down a track. In its path are five people who have been tied to the track by a mad philosopher. Fortunately, you could flip a switch, which will lead the trolley down a different track to safety. Unfortunately, there is a single person tied to that track. Should you flip the switch or do nothing?"

I must admit that when I first saw the trolley problem and subjected it to an economic analysis I was surprised that it had been given such a high regard, as it hardly seemed to me to be much of an ethical dilemma at all.  Using the ‘veil of ignorance’ principle we get a much clearer reason why – you don’t know any other details from behind the veil, only that you have a 50% chance of one person dying and a 50% chance of five people dying.  But it's much easier than that, because the trolley problem isn’t asserting that you have a 50-50 coin tossing situation and that you’ve got to hope the coin lands on the one person and not the five; it is saying that you can choose the outcome, which, to me, makes it a no-brainer.  It seems fairly obvious to me that in a straight choice between five people dying and one person dying we should choose the one.  Under Harsanyi’s principle we should choose the one to die because if you have amnesia and cannot remember whether you are the one or one of the five, probability favours your being one of the five. 

The next variation on the trolley problem is, instead of flicking a switch you have to push a heavy bystander on the track to stop the trolley, killing him and saving the five into which the trolley is heading.  Surprisingly, some people who were happy to flick the switch are less happy to push the heavy bystander onto the track.  On economic grounds there really should be no difference - if the only way to stop the trolley and save the five is to push one bystander on the bridge in front of the trolley we should do it, because, again, if you have amnesia and cannot remember whether you are the one on the bridge or one of the five on the track, probability favours your being one of the five. 

I said in the last Blog that duty and good intentions are high on our list of strivings, but there are also cases that bring us into direct contact with something potentially conflicting.  The philosopher Frances Kamm called this 'The Principle of Permissible Harm', which is the notion that one may endorse some forms of harm occurring but only if such harm is an effect or an aspect of the greater good itself. Some people claim that proponents of deontology do not have this option available to them, but this isn’t strictly true.  In the trolley situation the deontologist knows that some harm will come to someone (either one person or five people), so the option of ‘no harm’ isn’t available to him.  Not only does Harsanyi’s principle endorse flicking the switch to kill the one and save the five, Kamm’s ‘Principle of Permissible Harm’ says the same thing, because in Kammian terms of harm, you should push the bystander onto the track because if the five are killed this would be 5 contraventions of the Kamm’s Principle of Permissible Harm edict, whereas flicking the switch or throwing one innocent bystander in front of the trolley would only be 1 violation. 

You may object that there is a touch of self-interest in proposing the decision to kill one, but while that’s true, it doesn’t matter because, as we saw in the last Blog, probability favours self-interest and the collective good.  Here’s why it applies to the trolley problem.  If we consider Harsanyi’s principle again, not knowing which person (out of 6) you’d be in the situation, you are naturally hoping for the optimum outcome (as are the other 5).  You know you are one of the 6, so with the amnesia principle you know you are five times more likely to survive if you flick the switch or push the bystander.  It is ironic that most people’s protestations appear to be against the factor of self-interest, because here’s what they’re missing – the self-interest is not just self-interest for one, it is self-interest for all involved because when the probability is maximised for one it is maximised for everyone in the scenario.

And aside from self-interest we have beneficence too; for as long as we are behind the veil of ignorance we have the ideal balance of self interest (I want my own chances of death to be minimised) and beneficence (as long as I am choosing optimally, I am guaranteeing the best interests of the sum of people). For that reason I find it peculiar that the majority of people say they would flick the switch in the first example yet not push the bystander off the edge in the second.  In case anyone's still unsure, let me give a recap to show why an economic analysis doesn't support that view.  In Harsanyi's principle the scenario consists of the person standing on the edge, and five others at the end of the track about to die.  You don't know where you'll be in this model, which is why you should opt for the person standing on the edge to be pushed because, before you know where you'll fit, each of you has a 5/6 chance of being safe from death and a 1/6 chance of dying.  So before any information has been revealed, even the one who will end up being pushed still should have been in favour of one of the six being pushed because it gives him a 5/6 chance of being one of the safe members in the scenario (that’s why self-interest for you is simultaneously self-interest for everyone else in the scenario).  In opting for the best decision for himself he also opts for the best decision for the group, and in opting for the best decision for the group he also opts for the best decision for himself.

Now interestingly, I have a feeling that if we bring in a more emotive consideration things will change again.  I think this will give you pause for thought.  Suppose that instead of pushing the heavy bystander on his own, it emerges that his weight will only stop the trolley if your weight is combined too.  In other words, to save the five people you have to push him off and jump with him.  I have a feeling that more people will approve of that scenario than if the heavy bystander was pushed by himself.  The reason being, to jump with the heavy bystander suggests an instance of self-sacrifice on your part - a sacrifice that evidently must appear to have the best interests of the majority (the five at the end of the track) at heart, because it is still a 5/2 ratio of surviving members as opposed to the much worse 2/5 ratio. 

But this is where I think things become peculiar, because to me it seems clear that anyone who prefers you and the heavy bystander landing on the track over just the heavy bystander landing on the track probably does not have a very good grasp of moral dilemmas and is definitely not a very good economist.  Everyone should want to live in a world in which a 5/1 or 5/2 ratio of surviving members is chosen from behind the veil over a 1/5 or 2/5 ratio of surviving members.

Sometimes the moral philosopher pushes the situation further by asking if we’d willingly take the organs of the one man (killing the one) in order to save five other who needed his organs (enabling the five to survive).  Most people who think flicking the switch and pushing the fat man are ok draw the line at this third proposal, claiming the act of dismembering to be a little too extreme.  They have reached the right conclusion (I wouldn’t favour dismembering the one person either) but I suspect for the wrong reasons.  Maybe it takes an economist to conclude what a philosopher is less likely to conclude.  It isn’t because it is ‘too extreme’ that we should resist doing it - it is because the overall societal costs would outweigh the benefits.  Under Harsanyi’s principle nothing has changed with regards the probability – in economic terms you should still choose to kill the one, because from behind the veil of ignorance you and everyone else in the scenario are five times more likely to survive if you dismember the one man and use his organs to save the five.  But here’s why we should want to live in a world in which resisting dismemberment is chosen from behind the veil and not dismemberment.  Remember that from behind the veil of ignorance we are looking to choose a society that is optimal, not knowing where we’d fit into that society.  The reason we should resist is because we certainly would not wish to create a society in which people used our organs against our will to save others, because that would amount to a terribly unhealthy population full of individuals that didn’t pay much regard to looking after their body, and had much less incentive to do so.  If I know that any time soon I could be dismembered against my will to save the lives of others I’d have no qualms about living an excessive and unhealthy lifestyle - and the same would go for the rest of the population.

In actual fact, that is a good general rule of thumb regarding why utilitarianism by itself isn’t a strong enough foundation for an optimally healthy society.  Those who are opting for the utilitarian choice usually decide that it’s better to save the most and kill the fewest in any given scenario.  That’s why to a strict utilitarian saving the lives of five patients in a hospital by taking the organs of an unwilling sixth member and causing his death appears to be justified on utilitarian grounds, because it seems to maximise utility by sacrificing one person to save five.  The difficulty is, most people find the idea of using people as means to the end of promoting utility abominable if there is no recourse to moral duty.  I find I agree with them – it is questionable because on this instance it draws attention to the wider societal harms caused by actions such as this.  Strict utilitarianism fails because it would be horrible living in a land in which one might, at any moment, be sacrificed or compromised for the greater good.  Those who lived under such a regime could not be happy very easily, because they would more likely be neurotic, forever afraid of that knock on the door from the utilitarian Government’s officers. As such, actions which flout the rights of individuals (such as the transplant case) lead to a decrease in utility.

Regarding the different scenarios in the trolley problem, as far as psychology goes, it is perfectly reasonable to treat each case with different levels of revulsion.  I would always feel less uncomfortable flicking a switch than I would pushing a heavy bystander off a bridge.  I would always feel less uncomfortable pushing a heavy bystander off a bridge than I would harvesting the organs of an innocent man.  I almost certainly would feel inclined to kill the five and save the one if the one was my beloved, my child or a parent.  These feelings are perfectly natural – we have evolved to value those close to us more than strangers, and we have evolved to be repulsed at murder irrespective of whether there are five to one survival ratios that would favour such an act.

I’m not 100% sure what it says about human minds that so many people would flick a switch to save five and kill one but not push a heavy bystander off the edge to do the same, and that a situation with a 2/5 death ratio where the pusher self-sacrifices along with the man pushed is often preferred to a 1/5 death ratio where the man pushed is the only one that dies.  Maybe it just shows that these kinds of human moral instincts are misleading us.  Or maybe it shows that becoming too personally involved distorts our consequentialist analysis.  For example, if instead of physically getting up close and pushing the heavy bystander we had to flick a switch that dropped him in front of the trolley saving the five, and the switch was in a room one mile away next to a live videocam of the five people at the end of the track, we might well feel differently.

Finally, despite economic analysis to the contrary, I’m also unsure why it feels better and more honourable to grab the heavy bystander’s hand and jump with him rather than create a situation in which he falls by himself (and in doing so halves the death ratio). This is particularly strange, given that from behind a veil of ignorance all parties concerned would opt for the latter scenario.  Perhaps it shows that once we strip away all the things that pervert or confound us, we have much more of a subliminal likeness for self-sacrifice, love, grace, kindness and togetherness than we are often willing to admit.