What Is the Trolley Problem and What It Reveals About Moral Reasoning

The Runaway Trolley

A trolley is barreling down a track, its brakes failed, and five people are tied to the rails ahead. You are standing next to a lever. If you pull it, the trolley will be diverted to a side track — where one person is tied. Do nothing and five people die. Pull the lever and you actively cause one person's death. What do you do?

This is the trolley problem, introduced by philosopher Philippa Foot in 1967 and later developed by Judith Jarvis Thomson. It has become one of the most discussed thought experiments in the history of philosophy — not because it describes a realistic scenario, but because it forces a confrontation between two powerful moral intuitions that usually agree but here starkly conflict. Most people, when surveyed, say they would pull the lever: one death is better than five. This seems like straightforward math. But then Thomson introduced a variant that changes everything.

The Footbridge Variant

Imagine instead that you are on a footbridge above the same runaway trolley. Next to you stands a large man whose body, you somehow know with certainty, would stop the trolley if he fell onto the tracks. The only way to save five people is to push him off the bridge. Would you do it?

The arithmetic is identical — one death versus five deaths. Yet almost everyone who said "yes" to pulling the lever says "no" to pushing the man. This asymmetry is the heart of the trolley problem's philosophical power. Why does the same utilitarian calculation feel right in one case and deeply wrong in another? Something about using a person as a means — physically pushing them to their death — violates a moral intuition so strong that most people refuse even when the numbers favor it.

What the Trolley Problem Tests

Philosophers use the trolley problem and its variants to probe the structure of moral reasoning. The standard lever case appears to favor consequentialism — the view that the morally right action is the one that produces the best consequences. Five lives saved outweigh one life lost; pull the lever. But the footbridge case shows that most people are not pure consequentialists. The method of causing harm matters, not just the outcome.

This resistance is often analyzed through the lens of deontological ethics, particularly Kant's principle that persons must be treated as ends in themselves, never merely as means. Pushing the man uses his body as a trolley-stopping instrument — treating him as a tool for others' survival. There is also a relevant distinction, studied by moral psychologists, between doing harm and allowing harm: pulling the lever redirects an existing threat, while pushing the man initiates a new causal chain that involves directly harming someone.

The Psychology Behind the Intuition

In the early 2000s, neuroscientist Joshua Greene and philosopher Jonathan Cohen used brain imaging to study what happens when people reason through trolley variants. They found that the footbridge case, unlike the lever case, activates brain regions associated with emotional processing — particularly areas linked to personal, physical harm. The lever case is more emotionally distant; it feels more like an abstract problem.

Greene argued from this that our deontological intuitions are emotional responses shaped by evolution for face-to-face social environments, while utilitarian reasoning is more detached and deliberate. He controversially suggested that when emotion and calculation conflict, we should generally trust the calculation. Others pushed back hard: our moral intuitions are not just evolutionary accidents — they encode accumulated wisdom about how to treat persons, and overriding them based on abstract calculations is precisely how atrocities happen.

Variants That Deepen the Puzzle

Philosophers have created dozens of trolley variants to isolate specific moral features. A few of the most instructive:

• The loop track variant: The side track loops back to the main track, so the trolley would still hit the five — unless it struck the one person on the side track and stopped. Here, the one person is used as a means even in the lever case. Does that change your answer?
• The transplant case: A surgeon has five patients dying from organ failure. A healthy visitor has the right organs to save all five. Should the surgeon kill the visitor and harvest their organs? Most say no — but the arithmetic is the same as the footbridge case. This helps show that personal violence, not mere calculation, is what triggers deontological refusal.
• The fat villain variant: What if the large man on the footbridge were the person responsible for the trolley's brake failure? Does culpability change your willingness to push him? Many people say yes — punishment shifts the moral calculation.

Cross-Cultural Findings

Researchers have administered trolley problem surveys across dozens of cultures and found the same basic pattern — most people pull the lever in the standard case, most refuse to push in the footbridge case — with some variation in margins. This cross-cultural consistency is itself philosophically significant. It suggests the underlying moral intuitions are not purely culturally constructed but may reflect something more universal about how humans respond to harm, proximity, and using persons as instruments.

However, cultural context does shape responses. Studies have found variation in how much weight different populations give to group membership, authority figures, and the specific relationship between the decision-maker and the potential victims. The trolley problem isolates variables from real-world complexity, which is both its strength as an analytical tool and the source of legitimate criticism that it oversimplifies moral life.

Real-World Applications

The trolley problem is not purely academic. It directly informs debates about autonomous vehicle ethics: how should a self-driving car be programmed to respond when an accident is unavoidable — prioritize passengers, minimize total casualties, or avoid active harm to non-passengers? It appears in discussions of wartime targeting decisions, triage medicine, and public health policy where population-level benefits must be weighed against harms to individuals.

The MIT Media Lab ran a large-scale online experiment called the Moral Machine that gathered millions of judgments from people worldwide about self-driving car decisions. The results showed consistent cross-cultural preferences: spare more lives over fewer, prefer humans over animals, prefer the young over the old — but also cultural variation in how much these preferences were weighted. The data is being used to inform policy discussions about autonomous vehicle standards.

Why the Problem Endures

The trolley problem endures because it crystallizes a genuine tension in human moral psychology — one that cannot be dissolved by simply picking a side. Pure consequentialism, consistently applied, licenses horrifying conclusions (harvest the visitor's organs). Pure deontological refusal to violate persons, consistently applied, means watching five people die when you could save them. Real moral wisdom likely requires holding both intuitions, using each to check the excesses of the other, and reasoning carefully about which applies more strongly in any given situation. That is a much harder task than pulling a lever — which is exactly the point.