Thought experiment 10: The Trolley Problem

New Statesman​

23-04-2025

Illustration by Marie Montocchio / Ikon Images
Philippa Foot set the trains, or trolleys, in motion. In 1967, Foot, a well-to-do Englishwoman, the granddaughter of the former US president Grover Cleveland, published an article about abortion. Foot was then Oxford-based and this was the year that abortion was legalised in the UK. Among the questions that interested Foot was whether it might be right, or at least acceptable, to take the life of the foetus as a side effect of a medical procedure that was required to save the mother.
A variation of Foot's trolley case runs as follows. A train is hurtling down a track, its brakes having failed. Five people are tied with ropes to the track ahead. They will die if hit by the train. You are standing by the side of the track and could pull a lever diverting the train down a spur. The decision would be easy were it not for the fact that one person is tied on this spur. Although Foot's explanation is contentious, most people agree with her that it would be right to pull the lever.
In 1987 an American, Judith Jarvis Thomson (whom we've met in this column before – see Thought Experiment 2: The Unconscious Violinist) entered the debate with a variation: Footbridge. Once again, the out-of-control train is hurtling towards five innocents. This time you're on a footbridge next to a fat man (to appease modern sensibilities, now portrayed as a man carrying a heavy backpack). If you push the man over the bridge to his death, his weight will stop the train.
The puzzle is this: in both cases it looks like the option is to kill one to save five, but in Footbridge, most people believe it would be wrong to topple a man to his death. Why the difference? This is the Trolley Problem, or what Thomson called 'a lovely, nasty difficulty'.
It has sparked a mini-industry ('trolley-ology') of academic papers about runaway trains; in some of them, the scenario is so ludicrously convoluted that it is hard to see why we should give credence to any intuitions they elicit. But the problem has proved not merely to be of interest to philosophers. Psychologists and neuroscientists have jumped on the tram-wagon; so too have political scientists and sociologists.
Here are the sorts of things that have been studied. Does it make a difference if you're forced to wait a few seconds before responding to trolley scenarios, or if the scenarios are presented in your second language? What happens if those who'll be killed are in your 'out-group'? Suppose you could push one chimpanzee over the footbridge to save five chimpanzees – in other words, is it more acceptable to apply a crude utilitarian calculation to animals than humans?
My favourite study varied the name of the Fat Man. He was called Tyrone Paton (a stereotypical African American name) to one set of subjects and Chip Ellsworth III (a name redolent of old white money) in another. Pushing Tyrone would save 100 members of the New York Philharmonic. Pushing Chip would save 100 members of the Harlem Jazz Orchestra. Conservatives, the study claimed, were indifferent between these choices, but liberals (presumably keen to demonstrate their anti-racist credentials) were more likely to push poor old Chip.
Subscribe to The New Statesman today from only £8.99 per month Subscribe
Unusually for a philosophical thought experiment, the Trolley Problem has escaped the confines of the ivory tower, becoming a meme on social media, even turning up in TV programmes, such as The Good Place.
My own view is that the Trolley Problem has a solution – or at least an explanation. The reason it's acceptable to turn the train down the spur is that you don't intend the death of the person tied there. If this person were to escape from the ropes, and flee in time, you'd be delighted. Then nobody would die. But you need the death of Backpack Man – for if he were somehow to bounce off the track and run away, the train would trundle on and kill the five.
There are implications here for many areas of life, including the conduct of warfare. There's an important moral distinction between 'intending' to kill civilians and a legitimate military operation in which it's foreseen that some civilians will die. For many years now, the Trolley Problem has been taught to future officers training at the US military academy West Point.
There are potential implications too for autonomous vehicles. A driverless car might be unable to break in time from an accident immediately ahead of it, and face a 'choice' to plough on or to veer to one side. Perhaps on the left there are two children, and on the right three adults. How should the car be programmed to respond?
Still, trolley-ology, one of the most famous thought experiments, has been so exhaustively analysed one detects a sense of exasperation within the profession whenever the topic is wheeled on. The out-of-control train may finally be running out of steam.
David Edmonds is the author of 'Would You Kill the Fat Man?'.
[See also: Thought experiment 9: Mary's Room]
Related