The Trolley Problem: Ethics, Variants, and Why It Matters
The trolley problem asks whether to divert a runaway trolley to kill one person instead of five. Explore its variants, the ethical theories, and AI applications.
The trolley problem is a thought experiment in ethics. A runaway trolley is heading toward five people tied to the track. You stand beside a lever that can divert it onto a side track, where it would kill one person instead. Should you pull the lever? Most people say yes. But change the setup, so that you must push a large stranger off a bridge to stop the trolley, and most people say no, even though the arithmetic is the same. That gap has driven decades of research in philosophy, psychology, and neuroscience.
Where does the trolley problem come from?
The British philosopher Philippa Foot introduced the lever case in 1967, in an essay about abortion and the doctrine of double effect. Judith Jarvis Thomson, an American philosopher, gave it the name "the trolley problem," added the footbridge variant in 1976, and kept developing it in 1985. Others produced further versions: a loop variant in which the side track curves back to the main line, so that the one person's body is what stops the trolley, and a case in which the man on the bridge is the villain who tied the victims to the track. The variants matter because each one changes a single feature of the situation, which lets philosophers ask which feature actually drives our judgments.
What do people actually say?
In large online surveys, such as the Moral Sense Test run by Marc Hauser and colleagues and reported in 2007, around nine in ten respondents approved of diverting the trolley, while only about one in ten approved of pushing the man off the bridge. Answers vary somewhat between cultures and age groups, but the basic pattern is widespread and appears in many languages.
How do ethical theories answer it?
| Theory | Lever case | Footbridge case | Core reasoning |
|---|---|---|---|
| Utilitarianism | Pull | Push | Outcomes decide: five lives outweigh one in both cases |
| Deontology | Often permitted | Forbidden | Persons may not be used as mere means; doing differs from allowing |
| Doctrine of double effect | Permitted | Forbidden | Harm as a side effect may be allowed, harm as a means may not |
| Virtue ethics | Depends | Depends | What would a person of good character do? |
A utilitarian says the number of lives saved decides everything, so the right act is to pull the lever and also to push the man. A deontologist says some acts are off-limits whatever their results, and that using a person as a tool to save others violates their rights. The doctrine of double effect, which comes from Thomas Aquinas, says that harming someone as a foreseen side effect can be acceptable, but harming someone as the means to a goal is not. In the lever case the one person's death is a side effect of redirecting the trolley; in the footbridge case the man's death is how the trolley is stopped.
What does the brain do?
In 2001 Joshua Greene and colleagues used brain scans to show that the personal, hands-on footbridge dilemma activated areas linked to emotion more strongly than the impersonal lever dilemma. Greene proposed a dual-process theory: a fast, automatic emotional response that resists directly harming someone competes with slower, deliberate reasoning that adds up the outcomes. Whether the emotional response is a piece of moral wisdom or a bias is still debated. The way a dilemma is worded also changes the answers, which is an example of the framing effect.
Does the trolley problem matter in real life?
Versions of it appear in medical triage, military ethics, and the design of autonomous vehicles. In the Moral Machine experiment, run by the MIT Media Lab and reported in 2018, millions of people in 233 countries and territories made about 40 million decisions on how a self-driving car should act in unavoidable crashes. People broadly preferred sparing more lives over fewer, younger over older, and humans over animals, with notable differences between regions in how strongly they held these preferences.
Engineers point out that real cars almost never face a clean choice of this kind: the practical aim is to brake early and avoid the crash altogether. Critics also note that the thought experiment strips away context, such as uncertainty, relationships, and responsibility. But that is exactly its purpose. Like a laboratory experiment, it holds everything constant except one feature so that we can see which intuitions are doing the work, and then decide whether we trust them.
Tags
decision making ethics moral philosophy thought experiments trolley problem