The Trolley Problem

Philippa Foot's original question

Philippa Foot introduced the trolley in 1967, not as a standalone puzzle but as part of an essay on abortion and the doctrine of double effect. She needed a case that would isolate a specific moral distinction: is there a difference between killing and letting die? Between intended harm and harm that is merely foreseen?

The trolley was designed to produce the intuition that yes, these distinctions matter, even when the arithmetic is identical. Five people die if you do nothing. One person dies if you act. Foot's point was that this doesn't settle the moral question. How you bring about an outcome matters, not just what the outcome is.

What's actually being tested

Most people say they'd pull the lever. Five lives outweigh one. The math seems obvious. But Judith Jarvis Thomson introduced a variation in 1985: instead of a lever, you're on a footbridge above the tracks. The only way to stop the trolley is to push a large stranger off the bridge, using his body as a brake. Same arithmetic, but almost nobody says they'd do it.

The trolley problem isolates this discomfort. Our moral intuitions aren't simply utilitarian. We treat actively harming someone differently from redirecting harm, even when the outcomes are identical. The question is whether that difference is principled or just a psychological quirk we've mistaken for a moral principle.

The doctrine of double effect

The doctrine of double effect, a principle with roots in Thomas Aquinas, offers one way to explain the difference. It holds that causing harm as a foreseen side effect of bringing about a good outcome can be permissible, while causing the same harm as the means to that outcome is not.

In the lever case, the one person's death is a side effect of diverting the trolley. You're not using their death to save the five. In the footbridge case, the death of the man you push is your means. You're saving five by using his body as a brake. The doctrine says that's what makes it wrong.

Critics point out that this distinction can seem arbitrary. If you know pushing the man will kill him, and you push him anyway, how is his death a "side effect" rather than an intention? The doctrine requires a precise account of what you intend versus what you merely foresee, and that line is harder to draw than it looks.

Consequentialism and Kantian ethics

The trolley problem is a stress test for two dominant frameworks in moral philosophy.

Consequentialism, associated with Jeremy Bentham and John Stuart Mill, holds that the right action is the one that produces the best overall outcome. Pull the lever. Push the man. Five lives saved is better than one. The identity of the person who dies is morally irrelevant. What matters is the aggregate.

Kantian ethics disagrees. Kant's categorical imperative forbids treating persons merely as means to ends. Pushing a man to his death, even to save five, treats him as a tool. His rights impose a constraint that arithmetic can't override.

Most people's intuitions split between these frameworks depending on the case. They'd pull the lever but won't push the man. The trolley problem is useful precisely because it forces you to notice that you're not consistently committed to either view.

The loop case

Thomson also introduced the loop variant, less well known but philosophically sharper. Imagine the side track loops back toward the five people. Diverting the trolley will only stop it if it hits the one person on the loop. Their body stops it. If they weren't there, the five would still die.

In this case, the one person's death is not a side effect. It's load-bearing. The doctrine of double effect no longer clearly permits diverting. Yet most people still say they'd pull the lever.

This suggests our intuitions aren't tracking the doctrine as precisely as the doctrine implies they should. Something else is doing the work: physical distance from the harm, the directness of the action, something about using a lever versus touching a person. The loop case makes it harder to say exactly what.

What does neuroscience say about trolley problem decisions?

In the early 2000s, psychologist Joshua Greene used brain imaging to study how people respond to trolley variants. He found that personal cases like the footbridge activate regions associated with emotional processing more strongly than impersonal cases like the lever. His interpretation: the footbridge feels wrong because pushing someone feels wrong at a visceral level, independent of any moral reasoning.

Greene argued this supports a broadly consequentialist view. Our deontological intuitions aren't tracking moral truth. They're tracking disgust and proximity. Dispassionate reason recommends the utilitarian answer; evolution has wired us to feel the deontological one.

Critics pushed back hard. The fact that an intuition has a causal explanation doesn't mean it's wrong. We have evolutionary explanations for why we believe in logical consistency, and that doesn't make logic unreliable.

Real-world versions

The trolley problem shows up, slightly disguised, in decisions people actually make.

Self-driving car manufacturers faced it directly when programming collision-avoidance systems. If a crash is unavoidable, should the car prioritize the passenger or minimize total harm? A 2016 survey found that people morally approved of utilitarian self-driving cars but preferred to own one programmed to protect its own passengers. Consequentialism for other people's cars.

Medical triage during mass casualty events follows similar logic. Treating patients most likely to survive means not treating others who might have survived with more resources. Wartime rules of engagement involve comparable calculations. The trolley problem is abstract by design. The structure it exposes is not.

Would you pull the lever? And if your answer changes when you're standing on the bridge, what does that tell you about the principles you thought you held?