The Trolley Problem and Why It Still Matters

The trolley problem is a thought experiment that has become a cultural reference point. It is taught in introductory ethics courses, discussed on Twitter, parodied in The Good Place, and increasingly invoked in debates about self-driving cars and medical triage. Behind the meme is one of the most-discussed problems in moral philosophy of the last fifty years.

This article covers the classic formulation, the major variations, what philosophers actually argue (it's more interesting than the popular discussion), the real-world applications, and the substantial critique of trolley problems as a way of thinking about ethics.

The classic formulation

The trolley problem was first formulated by Philippa Foot in 1967 and elaborated by Judith Jarvis Thomson in 1976.

The standard case (Switch):

A runaway trolley is heading toward five people tied to the track. You are standing next to a switch. If you flip the switch, the trolley will divert onto a side track, where one person is tied. The five live; the one dies. If you do nothing, the five die.

Should you flip the switch?

Most people say yes. Saving five at the cost of one seems like the right trade. The utilitarian arithmetic supports this: 5 > 1, fewer deaths is better.

The classic variant

The footbridge case (also called Fat Man):

The same runaway trolley is heading toward five people. You are standing on a footbridge over the tracks. Next to you is a large person. If you push them off the bridge, their body will stop the trolley, saving the five. The five live; the one dies.

Should you push the person?

Most people say no. Even though the utilitarian arithmetic is the same (one death for five), the action feels different. Pushing someone to their death is not the same as flipping a switch.

The puzzle: why do these two cases produce different intuitions, when the outcome is the same?

This is the trolley problem in its philosophical interesting form. It's not "what should you do?" — it's "why do we treat these cases differently, and is that distinction defensible?"

What philosophers actually argue

The serious philosophical literature on the trolley problem has explored several distinctions that might explain the intuitive difference between Switch and Footbridge.

Distinction 1 — Action vs allowing.

In Switch, you redirect harm. In Footbridge, you use a person as a means. Even if the outcomes are the same, the moral structure may differ — direct use of a person as instrument is more morally weighty than redirection of an existing harm.

This is roughly the Kantian view: using a person as a means is wrong even if the outcome is good. Kant's categorical imperative requires treating persons as ends, not merely as means.

Distinction 2 — Doctrine of Double Effect.

A long Catholic tradition (originating with Aquinas) distinguishes effects that are intended from effects that are foreseen but not intended. In Switch, the one death is foreseen but not intended (it's a side effect of redirecting the trolley); in Footbridge, the death is the means (you push the person specifically to stop the trolley with their body).

Under the doctrine of double effect, the two are morally different — even though both produce the same outcome.

Distinction 3 — Trolley/Switch dilemmas occur naturally; Footbridge dilemmas don't.

Some philosophers (notably Judith Jarvis Thomson in her later work) have argued that our intuitions about Switch may be reliable because they reflect real moral structure — but our intuitions about Footbridge may be unreliable because the situation is too artificial to map to real-world ethics. The cases feel different because Footbridge invokes our visceral aversion to direct violence, which may or may not track genuine moral truth.

Distinction 4 — Personal force.

Joshua Greene (Harvard) used fMRI studies to show that Footbridge-style cases (involving personal force) activate emotional brain regions; Switch-style cases (involving abstract instruments) activate more deliberative regions. The intuitive difference may reflect cognitive processing, not moral principle.

This raises a separate question: even if the intuitive difference is just emotional, does that mean we should disregard it? Some philosophers (Peter Singer, the utilitarians) say yes; others say our emotional responses encode wisdom that conscious reasoning misses.

The variations multiply

Beyond Switch and Footbridge, the literature has produced many variations.

The doctor: A doctor has five patients who will die without organ transplants. A healthy person walks in for a checkup. Should the doctor kill the healthy person to save the five?

This is structurally similar to Footbridge — using a person as means. Nearly everyone says no.

The loop: A variation where the side track loops back to the main track, but the one person's body would stop the trolley before it loops back. This makes the harm to the one person essential (not a side effect). Does this change the moral analysis?

The bystander at the switch: A variation where you're a bystander rather than a designated operator. Does that change your obligations?

The asymmetric numbers: 100 vs 1. 5 vs 5. 5 vs 4. At what point do the numbers start to matter, and at what point do they stop mattering?

Each variation probes a different aspect of moral intuition. The cumulative effect: the trolley problem is not really a single problem; it's a family of problems that test different ethical principles.

Real-world applications

Some philosophers complain that trolley problems are too artificial to inform real ethics. The complaint has weakened in recent years as actual trolley-like cases have appeared.

Self-driving cars.

Suppose an autonomous vehicle must choose between protecting its passenger and avoiding pedestrians. How should it choose? Under what circumstances does it choose the pedestrian's life over the passenger's? These are trolley problems, and engineers must answer them in code.

MIT's Moral Machine project crowdsourced responses to autonomous-vehicle dilemmas. The results showed substantial cultural variation in what people think the right answer is — which itself is consequential.

Medical triage.

In a mass casualty event, doctors must allocate scarce resources (ventilators, surgeries, ICU beds). They explicitly trade some patients' survival for others'. This is trolley-problem reasoning in real life.

Pandemic response.

Public health policies during COVID-19 — lockdowns, mask mandates, vaccine prioritization — required tradeoffs between lives, livelihoods, freedoms, and other values. The trolley problem framework illuminates the structure of these decisions, even if the actual decisions were more complex.

Military and policing.

Drone strikes, hostage situations, and use-of-force decisions all involve dilemmas where some people are killed to save others. The trolley problem is the simplified, ethically pure version of these messy real-world choices.

Climate policy.

Climate change involves tradeoffs across generations: current quality of life vs. future climate stability. This is a temporal trolley problem.

The critique of trolley problems

Several philosophers have argued that trolley problems are misleading. The strongest version:

Critique 1 — They abstract away the things that matter.

Real ethical situations involve uncertainty, relationships, history, institutions, and the texture of lives that trolley problems strip out. The artificial purity of the problems makes them tractable for analysis but unrepresentative of actual ethics.

Critique 2 — They privilege a narrow type of decision.

Trolley problems are act-focused: what should you do at this moment? But much of ethics is character-focused (what kind of person should you be?) or institutional (what should the rules be?). Trolley problems systematically underemphasize these dimensions.

Critique 3 — They invite false confidence.

The simplicity of the trolley problem makes ethical principles feel cleaner than they actually are. People who have read trolley problems sometimes apply them to real situations with more confidence than is warranted.

Critique 4 — They may produce wrong answers in real life.

Some scholars (Bernard Williams, Martha Nussbaum) have argued that pure utilitarian reasoning, which trolley problems often endorse, produces conclusions that conflict with strongly-held moral intuitions in real cases. The trolley problem may train us into a moral framework that fails in actual circumstances.

The defense: trolley problems are not meant to be road maps to real ethics. They're meant to illuminate the structure of ethical reasoning by isolating specific moral variables. The simplicity is the point. Applied to real situations, they need to be supplemented by other considerations — but they remain useful as analytic tools.