n.Newfeed

Philosophy · 6 min read

What the Trolley Problem Was Originally Asked to Explain

A return to Foot and Thomson’s original arguments and what later experiments can—and cannot—add.

What the Trolley Problem Was Originally Asked to Explain

The thought experiment before the meme

The trolley problem is often presented as a personality test: pull a lever and kill one person to save five, or refuse and let five die. Online versions count answers, assign moral labels, or replace the people with comic objects. That use obscures the problem’s original function. Philippa Foot introduced the runaway tram in a 1967 essay to examine whether the doctrine of double effect could explain contrasting judgments about harming people. Judith Jarvis Thomson later named and reformulated the “trolley problem” as a puzzle about why apparently similar cases elicit different moral verdicts.

The thought experiment was not designed to prove that lives should be added like numbers. Nor was it a prediction of how a real driver would behave under stress. It isolates selected features—agency, intention, rights, means, side effects, and duties—to test whether a proposed moral distinction holds across cases.

That makes the trolley case more like a counterexample in an argument than a miniature policy referendum. A response may reveal an intuition, but the philosophical work begins when one asks which difference between cases justifies it and whether that principle survives further examples.

Foot’s question about double effect

Foot’s essay, “The Problem of Abortion and the Doctrine of the Double Effect,” addressed a traditional claim: it can matter whether a harmful outcome is intended as a means or merely foreseen as a side effect. The doctrine had been used in discussions of medical treatment and war. Foot tested whether intention alone explained judgments across cases.

In one comparison, a driver of a runaway tram can steer from a track with five people onto a track with one. In another, a judge could frame and execute one innocent person to stop riots that would kill five. Many readers permit steering but reject the frame-up, even though each choice appears to trade one life for five. If numbers alone governed the answer, the cases should align.

Foot argued that distinctions among duties help. A negative duty not to injure can be more stringent than a positive duty to provide aid. The judge would deliberately violate the innocent person’s right in order to benefit others. The tram driver faces an existing lethal threat and redirects it while operating the vehicle. Foot’s full discussion is more cautious than the familiar lever vignette: she examines abortion, medical scarcity, rescue, and what counts as doing harm.

The doctrine of double effect did not simply settle the comparison. Foot questioned formulations that make intention carry too much weight. A person can describe an intended means narrowly to redescribe harm as a side effect. Conversely, foreseeing death with certainty does not make responsibility disappear. Her cases probed the doctrine rather than serving as an advertisement for it.

Thomson’s reformulation of the problem

Thomson’s 1976 essay asked why turning a trolley can appear permissible when apparently comparable killings do not. She varied the setup to remove role-specific explanations. In the now-familiar footbridge case, a bystander can stop the trolley only by pushing a large person into its path. Many who would turn the switch refuse to push.

The contrast challenges a simple rule of minimizing deaths. It also challenges intention as a complete answer. In both the switch and footbridge cases, the one person’s death may be foreseen; in footbridge, however, the person’s body is used as part of the means of stopping the trolley. Later discussion has examined whether that means–side-effect distinction, physical contact, personal force, rights, or the structure of the threat explains the judgment.

No variant is neutral. Saying that a person is “large,” “innocent,” “on a side track,” or “already threatened” changes features that may be morally relevant or emotionally salient. Philosophers create sequences of cases to hold some features constant while changing others. The method can expose inconsistency, but it can also generate disputes over whether the cases are genuinely parallel.

Thomson herself revised her views. In a 1985 essay she distinguished a bystander at a switch from an agent who would divert a threat toward someone else and later argued that rights constrain what may be done. The literature did not freeze after one lever. The enduring problem is to state a principle that explains considered judgments without producing unacceptable implications elsewhere.

What experimental responses add

Moral psychology turned trolley cases into experiments. Joshua Greene and colleagues’ 2001 functional-imaging study contrasted “personal” dilemmas, such as pushing, with more impersonal cases, such as switching. It found different patterns of response time and brain activity, contributing to a dual-process account in which emotional and controlled processes can compete.

The finding did not locate a utilitarian module or show that one answer is rational and the other emotional. Brain regions participate in multiple tasks, and the categories “personal” and “impersonal” bundle several features. Later researchers have altered wording, order, realism, perspective, and cultural context, finding that judgments are sensitive to framing and individual differences.

Large cross-cultural projects show both recurring tendencies and variation. People often distinguish direct instrumental harm from redirecting an existing threat, yet rates of approval differ across samples and scenarios. Online convenience samples are not populations, translated vignettes may not be equivalent, and choosing a button in a hypothetical case carries none of the uncertainty, bodily stress, or accountability of an emergency.

Experiments can reveal response patterns and test psychological mechanisms. They cannot by themselves determine which response is morally justified. Moving from “most participants chose the switch” to “switching is right” would infer a normative conclusion from a descriptive frequency. Conversely, evidence that emotion influences a judgment does not discredit it; concern for another person is both emotional and potentially morally relevant.

What the cases do not decide

Trolley cases simplify away institutional responsibility, prevention, uncertainty, and alternatives. Real transport safety asks why brakes failed, who designed the system, whether warnings work, and how risk should be distributed before an emergency. Autonomous-vehicle policy involves uncertain probabilities, legal duties, occupants, road users, and system-wide incentives. Programming “the trolley answer” would not resolve those questions.

The cases also do not divide people neatly into consequentialists and deontologists. A person may turn the trolley because the one is already within the same threat, because the driver has a special duty, or because redirecting is not using the person as a means. Someone may refuse to push because of rights, uncertainty, aversion to personal violence, or doubt that the body would stop the trolley. The same choice can follow different reasons.

What is established historically is that Foot used the tram within an argument about double effect and duties, and Thomson transformed the contrast into a named problem. What is established empirically is that responses vary systematically with features such as personal force and instrumental use, while also varying across people and presentations.

What remains contested is the best moral explanation. Intention, means, rights, agency, and consequences each capture part of the pattern and face counterexamples. The trolley problem’s durable contribution is not a verdict about five versus one. It is a disciplined demand to explain why a distinction matters—and to accept the consequences of that explanation in the next case.

Sources

4 sources, 1 of them with a link you can follow.

  1. Philippa Foot, “The Problem of Abortion and the Doctrine of the Double Effect,” Oxford Review 5 (1967): 5–15.
  2. Judith Jarvis Thomson, “Killing, Letting Die, and the Trolley Problem,” The Monist 59, no. 2 (1976): 204–217.
  3. Judith Jarvis Thomson, “The Trolley Problem,” Yale Law Journal 94, no. 6 (1985): 1395–1415.
  4. Joshua D. Greene and colleagues, “An fMRI Investigation of Emotional Engagement in Moral Judgment,” Science 293 (2001): 2105–2108, doi:10.1126/science.1062872.

Want the reader to remember this spot? Sign in — it is free.

Next Moral Luck and the Limits of Responsibility

The app

Keep going, or take it with you.

Free accounts read twenty pieces a day. Pro is US$1.99 a month or US$11.99 a year for unlimited reading.