The trolley problem
A runaway trolley will kill five people unless you divert it onto a track where it will kill one. Most people pull the lever. Change one detail, and most people refuse.
- Reading time
- 6 min
- Difficulty
- Introductory
- Branch
- Ethics
- First posed
- Philippa Foot, 1967; named and extended by Judith Jarvis Thomson, 1976 and 1985
The setup
A trolley is hurtling down a track toward five people who cannot get out of the way. You are standing next to a lever. If you pull it, the trolley will switch to a side track, where one person is standing. That person will be killed. If you do nothing, the five will die.
Most people, asked what they would do, say they would pull the lever. One death is better than five.
Now change the scene. You are on a footbridge above the track. The trolley is heading toward the five. Next to you stands a very large man. If you push him off the bridge onto the track, his body will stop the trolley. He will die, and the five will be saved. There is no other way; you are too light to stop it yourself.
The numbers are the same: one dies, five live. Yet most people say it would be wrong to push the man. This is the puzzle. If it is right to divert the trolley, why is it wrong to push the man? And if there is a morally relevant difference, what is it?
The original case was introduced by Philippa Foot in 1967, in a paper about abortion and the doctrine of double effect. Judith Jarvis Thomson named it 'the trolley problem' and invented the footbridge variant. It has since generated so many variations that some philosophers joke about 'trolleyology'.
Interactive
The trolley, six times
Six versions, one choice each. In every case, acting saves five lives at the cost of one. After each choice, you will see how people usually answer where studies exist.
Case 1 of 6 · The switch
A runaway trolley is heading for five people tied to the track. You are standing by a lever. Pull it, and the trolley switches to a side track, where it will kill one person instead.
What do you do?
In surveys
Choose one to see how people answer.
Case 2 of 6 · The footbridge
This time you are on a bridge over the track. The only way to stop the trolley is to push a large man off the bridge into its path. He will die; the five will live.
What do you do?
In surveys
Choose one to see how people answer.
Case 3 of 6 · The loop
Back at the lever. But the side track loops round and rejoins the main line. If the one person were not on it, the trolley would come back and kill the five anyway. His body is what stops it.
What do you do?
In surveys
Choose one to see how people answer.
Case 4 of 6 · The trapdoor
The man is on the bridge again, standing on a trapdoor. You are some distance away, at a lever that opens it. Pull it, and he drops in front of the trolley. No pushing; just a lever, like the first case.
What do you do?
Case 5 of 6 · The transplant
You are a surgeon. Five patients will die without organ transplants. A healthy visitor in the waiting room is a perfect match for all five. No one would ever find out.
What do you do?
Case 6 of 6 · Yourself
The first scene again, but now you are the one on the side track, and the lever is within your reach. You can divert the trolley onto yourself and save the five.
What do you do?
Your six answers
Why it matters
The trolley problem is a precise tool for testing moral theories. A pure consequentialist says both cases are the same, and you should push. Most people's intuitions disagree, which suggests either that consequentialism is incomplete, or that our intuitions are mistaken.
It is not only academic. Doctors deciding how to allocate scarce ventilators, governments weighing the costs of public health measures, and engineers designing the decision rules of self-driving cars all face structurally similar choices: whether, and how, it is acceptable to cause some harm to prevent greater harm.
Key terms
- Doctrine of double effectThe principle that it can be permissible to cause a harm as a foreseen side effect of pursuing a good end, when it would be wrong to cause the same harm as a means to that end.
- Doing and allowingThe distinction between causing a harm and merely letting it happen. Many people think doing harm is worse than allowing it, even when the outcome is the same.
- ConsequentialismThe family of moral theories holding that whether an act is right depends only on its consequences. Utilitarianism is the best-known member.
- DeontologyThe family of moral theories holding that some acts are required or forbidden in themselves, whatever their consequences, because of duties, rights or constraints.
- Means and side effectsThe distinction between harm you bring about as a way of achieving your goal and harm that is merely a foreseen by-product of achieving it. Central to the doctrine of double effect.
The positions
Push the man
The two cases are morally the same. In both, one dies and five are saved. If you should pull the lever, you should push the man.
Means versus side effect
Diverting the trolley is permissible because the one person's death is a foreseen side effect. Pushing the man is wrong because his death is the means by which the five are saved.
Doing versus redirecting
It is permissible to redirect an existing threat toward fewer people, but not to create a new threat or use a person as a tool.
Do nothing
You should not intervene in either case. Killing is worse than letting die, and you have no right to choose who dies.
Arguments and objections
Push the man
- Best argument
- What matters morally is outcomes: five lives are worth more than one. The difference in our reactions is a matter of psychology, not morality. Joshua Greene's brain-imaging research suggests that the footbridge case triggers strong emotional responses to up-close, personal violence, which evolved in small groups and may not track what is right. On reflection, we should trust the reasoning over the gut.
- Strongest objection
- This seems to license terrible acts whenever the numbers favour them, including killing a healthy patient to harvest organs for five others. Almost no one accepts that conclusion, so there must be something wrong with the reasoning.
- The reply
- Some consequentialists accept the hard cases; others argue that a policy of harvesting organs would destroy trust in medicine and so produce worse outcomes overall.
Means versus side effect
- Best argument
- The doctrine of double effect, traced to Thomas Aquinas, holds that it can be permissible to bring about a harm as a side effect of pursuing a good end, when it would not be permissible to bring about the same harm as a means to that end. In the lever case, you would be glad if the one person somehow escaped. In the footbridge case, you need the man to be hit; his body is your tool.
- Strongest objection
- Thomson's loop case: the side track loops back to the main line, so the trolley would hit the five from behind if the one person were not there. Now the one person's body is a means of stopping the trolley. Yet many people still think it is permissible to pull the lever.
- The reply
- Defenders argue that intuitions about the loop are divided, and that the doctrine captures something real even if it needs refinement.
Doing versus redirecting
- Best argument
- Thomson argued that in the lever case, you are turning an existing threat so that it falls on fewer people. In the footbridge case, you are introducing a new harm to someone who was not in the threat's path at all, and using their body. The distinction between doing and allowing, and between redirecting and initiating, matters morally.
- Strongest objection
- The distinctions multiply and become hard to state precisely. Thomson herself later changed her mind and argued that the bystander should not pull the lever, since you cannot ask the one person to bear a cost you would not bear yourself.
Do nothing
- Best argument
- There is a moral difference between killing someone and failing to save them. If you pull the lever, you kill the one; if you do nothing, the trolley kills the five, and you merely fail to prevent it. Many legal systems reflect this difference, and it protects people from being sacrificed whenever someone calculates that it would help others.
- Strongest objection
- It seems monstrous to let five people die when you could save them at the cost of one, especially when you are not using anyone as a tool. And doing nothing is itself a choice with consequences.
Who argued what
- Push the man Peter Singer, Joshua Greene, Jeremy Bentham
- Means versus side effect Thomas Aquinas, Elizabeth Anscombe, Philippa Foot
- Doing versus redirecting Judith Jarvis Thomson, Frances Kamm
- Do nothing Judith Jarvis Thomson (later view)
Where you stand
In the original case, do you pull the lever?
In surveys
Choose one to see how people answer.
Variations
The loop
The side track loops back to the main track. Without the one person on it, the trolley would circle around and kill the five anyway. His body is what stops it. Should you pull the lever? If so, you are using him as a means, which the doctrine of double effect seems to forbid.
The transplant
A surgeon has five patients who will die without organs, and a healthy visitor who is a perfect match. Almost everyone says killing the visitor would be monstrous. Foot used this case to argue that the duty not to harm is stronger than the duty to help.
The self-driving car
A self-driving car's brakes fail. It can swerve into one pedestrian or continue into three. How should it be programmed? The MIT Moral Machine experiment collected about 40 million decisions from people in 233 countries and territories and found broad agreement on sparing more lives and younger lives, with notable variation between cultures.
Go deeper
What the brain scans showed
In a 2001 study, Joshua Greene and colleagues found that 'personal' dilemmas like the footbridge case activated brain areas associated with emotion more strongly than 'impersonal' ones like the lever case. Greene developed a 'dual-process' theory: deontological judgments are driven by fast emotional responses, consequentialist ones by slower reasoning. The interpretation is contested, and the original findings have faced methodological criticism, but the research made the trolley problem a meeting place for philosophy and psychology.
Is the trolley problem useful?
Critics argue that trolley cases are so artificial, with certainties no one ever has, that they tell us little about real ethics. Defenders reply that this is the point: like a controlled experiment, a thought experiment isolates one variable at a time to see what makes a difference.
Schools of thought: Utilitarianism, Kantian ethics
Related ideas
Further reading and sources
- EncyclopediaStanford Encyclopedia of Philosophy: Doing vs. Allowing Harm
- WikipediaTrolley problem
- SurveyPhilPapers Survey, 2020 · question: Trolley problem; about 1,785 professional philosophers; accept or lean toward
- SurveyAwad et al., 'Universals and variations in moral decisions made in 42 countries', PNAS, 2020 · question: Switch; about 70,000 people in 42 countries; country-level averages
- BookWould You Kill the Fat Man?, David Edmonds
- BookThe Trolley Problem Mysteries, Frances Kamm
- BookMoral Tribes, Joshua Greene
Open questions
- What, if anything, is the difference between the lever and the footbridge?
- Should a self-driving car be programmed to sacrifice its passenger to save more people?
- Should you trust your gut reaction to these cases, or your reasoning?