The trolley problem

The tripartite distinction between doing, allowing and enabling harm, which I discussed in my last blog, can put many philosophical and daily problems in another light, for example the well-known trolley problem. Although the trolley problem is often discussed as a pure philosophical problem, there are many practical versions, like those treated by Dostoevsky and Sartre. Here I’ll focus on a philosophical version of the case, called Bystander: A driverless, runaway trolley on a railway is heading for a tunnel, in which it will kill five people, if nothing stops it. A bystander can save their lives by turning a switch and redirecting the trolley on to another track. However, there is a man walking on that track that will then be killed instead of the five.

There are also other versions of the philosophical trolley problem, for example that there is a driver on the trolley who can turn the switch; that a bystander can stop the trolley by pushing a fat man on the track; etc. But this simple version will do in order to show my point.

Judith Jarvis Thomson discusses the following principle that can guide the decision whether or not to turn the switch: “Though (1) killing five is worse than killing one, (2) killing one is worse than letting five die”.

The principle seems reasonable and moreover it is typical for the debate on the trolley problem, where actually every participant agrees that:

(I) the bipartite distinction between doing and allowing harm applies;

(II) turning the switch so that the single walker dies is an act of killing that the bystander does.

I start with assumption II. I have my doubts that turning the switch so that the single walker dies is an act of killing. Of course, it is so that the single walker is killed, if the bystander turns the switch, and it is also the case that killing is a matter of degree, ranging from killing by accident to outright murder. Nevertheless, if we say that the bystander kills the single walker by turning the switch, to me it sounds somewhat odd. It is as if the bystander has the intention to kill the single walker in order to save the five. However, the bystander doesn’t have this intention at all. He merely wants to save the five people. If he could save the single walker as well, he would be very happy. Moreover it is not the bystander who kills the single walker but it is the trolley that does. I think that there is more to say for it that the person who made that the trolley started moving or could move in the direction of the switch is responsible for the killing of either the five or of the single walker. If someone would be punished for the death of one or five persons respectively, it would be him.

Be that as it may, let me examine what the bystander does. The bystander has the intention to save the lives of the five persons on the track. The bystander thinks that he can save these five lives only by turning the switch. Therefore he turns the switch and the five persons are saved. However, the single walker is killed as an unintended consequence of this action. Should we say then nevertheless that the bystander kills the single walker? For, hadn’t the bystander turned the switch, the single walker hadn’t died. It looks the same as if the bystander had killed some by accident.

In order to solve this problem we must look at assumption I. The reason why we say that the bystander kills the single walker is that this assumption gives us only the possibilities to say either that the bystander does something (turning the switch) and then it follows that he kills the single walker; or that the bystander allows something to happen (the switch leave as it is) but then the single walker survives and the five people die. Since the bystander turned the switch, we must say that he killed the single walker. However, as we have seen in my blog last week, the bipartite distinction between doing and allowing harm is not correct and must be replaced by the tripartite distinction between doing, allowing and enabling harm. And if we do so, then we need no longer to say that the bystander killed the single walker, but that by turning the switch he enabled that the single walker was killed (namely by the trolley).

I think that introducing the category of enabling in the debate will put the trolley problem in a new light and maybe it will make that much reasoning on the theme must be revised. One thing that might be changed is the principle quoted above. I think that it’s true that (1) killing five is worse than killing one. It may also be arguabel that (2) killing one is worse than letting five die. However, now we should add a lemma like (3) letting five die is worse than enabling one being killed – assuming that it is possible to defend this view (which I am not yet sure of, since it depends a lot on the conditions that occur).

Reference: Judith Jarvis Thomson, “Kamm on the Trolley Problems”, in F.M. Kamm et al. The Trolley Problem. Mysteries. Oxford, etc.: Oxford University Press, 2016; pp. 113-134 (quotation on p. 114).