What is the right thing to do when every option costs lives? In the opening sessions of Harvard’s Justice course, Michael Sandel uses a runaway trolley, a bridge, a hospital, and a 19th-century shipwreck to show how ordinary moral instincts collide—and why the clash between consequentialist and categorical reasoning has shaped law, politics, and philosophy for centuries.
The Classic Trolley Problem: Five Lives vs. One
Sandel begins with a scenario: you are driving a trolley at 60 mph. Five workers are on the track ahead. The brakes fail. You cannot stop.
Then you notice a side track with one worker on it. You can turn the steering wheel, killing one person but sparing five.
When Sandel polls his audience in Sanders Theatre, the vast majority say they would turn onto the side track.
The most common reason: It can’t be right to kill five when you could kill one instead. One student compares it to the passengers on United Flight 93 on 9/11—tragic, but better to sacrifice fewer lives to save more.
Those who would not turn offer a very different warning. One student argues that “better to kill one to save five” is the same logic that can justify genocide and totalitarianism—sacrificing one group to preserve another. When pressed, she says she would still go straight and kill the five rather than adopt that principle.
The “Fat Man” Variant: Why Most People Change Their Answer
Sandel then changes the setup. You are an onlooker on a bridge—not the driver. A runaway trolley is heading toward five workers. Next to you stands a very large man. If you push him onto the track, he dies, but the trolley stops and the five survive.
This time, most people would not push.
That raises the central puzzle: what happened to the principle that one death is better than five?
Students offer competing explanations:
- Active vs. passive involvement — pushing someone who was not already in danger feels like choosing to involve an innocent third party
- Direct killing vs. redirecting harm — steering a runaway vehicle feels different from using your hands to kill someone
- Becoming involved as a bystander — in the first case, the driver is already part of the crisis; on the bridge, you can choose whether to enter it at all
When Sandel asks whether opening a trap door under the man—via a steering wheel rather than a push—would feel different, student Andrew still hesitates. Even indirect mechanisms can feel morally worse than turning the trolley in the original case.
Doctor Dilemmas: Triage and the Transplant Surgeon
Sandel moves the same logic into medicine.
Emergency room triage
Six patients arrive from a trolley wreck. Five have moderate injuries; one is severely injured. You can spend all day saving the one—and the five die—or treat the five and let the one die.
Almost everyone saves the five, using the same one-vs.-five calculus.
The transplant surgeon
Now you are a transplant surgeon. Five patients need a heart, lung, kidney, liver, and pancreas. No donors are available. In the next room, a healthy man is napping after a checkup. You could quietly remove his organs, killing him but saving five.
Almost no one would do it.
One student in the balcony tries to dodge the dilemma by suggesting harvesting organs from whichever of the five dying patients expires first—a clever workaround Sandel calls “a great idea” that “wrecked the philosophical point.”
The pattern is clear: when consequences alone seem to govern our choices in some cases, the character of the act itself stops us in others.
Consequentialist vs. Categorical Moral Reasoning
From these debates, two frameworks emerge.
| Framework | Core idea | Example in the lectures |
|---|---|---|
| Consequentialist | Morality depends on the results of an act—the state of the world it produces | Turn the trolley; save five at the cost of one |
| Categorical | Some acts are wrong in themselves, regardless of outcomes | Do not push the innocent man; do not harvest organs from a healthy patient |
The leading consequentialist tradition is utilitarianism, associated with 18th-century English philosopher Jeremy Bentham. The leading categorical tradition is associated with German philosopher Immanuel Kant.
Sandel previews the semester’s reading list—Aristotle, John Locke, Kant, John Stuart Mill—and contemporary debates on equality, affirmative action, free speech, same-sex marriage, and military conscription. The goal is not abstract trivia but understanding what is at stake in everyday political life.
Why Philosophy Is Unsettling—and Unavoidable
Sandel warns that studying justice carries personal and political risks.
Philosophy does not mainly supply new facts. It takes what we already believe in familiar settings and makes it strange. Once the familiar turns strange, it cannot fully return to innocence.
Politically, philosophy may even make you a worse citizen before a better one—because it distances you from conventions and settled assumptions. In Plato’s Gorgias, Callicles tells Socrates to quit arguing and “go to business school”—meaning: stop questioning and pursue reputation and success.
The characteristic evasion is skepticism: if Aristotle, Locke, Kant, and Mill never settled these questions, why should we?
Sandel’s reply: the questions persist because we live some answer to them every day. Kant wrote that skepticism is a resting place for reason, not a permanent home. The course aims to awaken “the restlessness of reason.”
Jeremy Bentham and Utilitarianism: “The Greatest Good for the Greatest Number”
In the next lecture, Sandel tests utilitarianism against a real case—not a hypothetical.
Bentham’s core claim: the right action maximizes utility—the balance of pleasure over pain, happiness over suffering. Humans are governed by two “sovereign masters”: pain and pleasure. Morality, law, and policy should maximize overall happiness.
The slogan: “The greatest good for the greatest number.”
Queen v. Dudley and Stephens: Cannibalism at Sea
The 1884 British case Queen v. Dudley and Stephens is a staple of law schools worldwide.
After the yacht Mignonette sank in the South Atlantic, four crew members escaped in a lifeboat: Captain Thomas Dudley, First Mate Edwin Stephens, sailor Brooks, and 17-year-old cabin boy Richard Parker—an orphan on his first long voyage.
They had two cans of turnips, a turtle, and then nothing for eight days. Parker drank seawater, grew ill, and appeared near death.
On day 19, Dudley proposed drawing lots to decide who would die. Brooks refused. No lottery was held.
The next day, Dudley told Brooks to look away, signaled Stephens, said a prayer, told Parker his time had come, and killed him with a penknife. All three survivors ate Parker’s body for four days until rescue on day 24.
Dudley’s diary entry on rescue day: “On the 24th day, as we were having our breakfast… a ship appeared at last.”
Back in England, Brooks turned state’s witness. Dudley and Stephens stood trial. They did not deny the facts. Their defense: necessity—better one should die so three could live. The prosecutor replied: murder is murder.
How the Class Judged Dudley and Stephens
When Sandel asks students to judge the act morally—not legally—a sizable majority say what Dudley and Stephens did was wrong. Roughly 20%, led by a student named Marcus, defend them.
Defense arguments:
- Extreme necessity after 19 days without food—“You’ve got to do what you’ve got to do”
- Survival may allow survivors to benefit society later (though Sandel presses: what if they became assassins?)
- Distinction between moral reprehensibility and legal guilt—necessity may reduce accountability even if the act is troubling
Prosecution arguments:
- Humans do not have the power to take others’ lives into their own hands—“We don’t have that kind of power” (Britt)
- If extreme hunger excuses killing, the same logic could excuse murder to feed a family on a street corner
- Murder is murder—even if Bentham’s math favors three lives plus dependents over one orphan
Consent, Lottery, and Due Process
Three refinements shift how many students judge the case—without fully resolving it.
Consent
Student Kathleen asks: what if Dudley had asked Parker, “Would you mind?” before killing him? Even if Parker murmured “okay” in a stupor, Kathleen still finds it unjustifiable.
Others say consent matters—but only if it is genuinely voluntary, ideally initiated by Parker himself, not coerced by three desperate men.
Lottery
If all four had agreed to a fair lottery—and Parker lost—more students say the act becomes morally permissible. For Matt, the core crime is deciding their lives matter more than his without a shared procedure. A lottery treats everyone as equally at risk.
Others stress Parker was never consulted about the lottery in the actual case—and had no forewarning that his life was on the line.
Categorical rejection regardless
Some students, including Mike, hold that murder is categorically wrong even to save 30, 300, or 3,000 lives. When Sandel invokes Bentham’s calculus—three sailors with wives and children vs. one orphan with no dependents—Mike replies that adding up collective happiness does not override the wrong of killing an innocent person. If Mike is right, Bentham must be wrong on this point.
Sandel closes by listing three questions the course will pursue:
- Why is murder categorically wrong—and do fundamental rights come from something other than utility?
- Why should agreement to a fair procedure justify its outcome?
- What moral work does consent actually do?
Bottom Line
The trolley problem is not a puzzle with one clever answer. It is a method for exposing how we reason about justice.
Most of us accept consequentialist logic when flipping a switch or triaging patients—but reject it when pushing a stranger off a bridge or harvesting organs from a healthy person. Real cases like Dudley and Stephens force the same tension: necessity, numbers, consent, and procedure all matter, yet many still insist some acts are wrong no matter the outcome.
Sandel’s opening lectures set up the semester’s central debate: Jeremy Bentham’s utilitarianism vs. Immanuel Kant’s categorical duties—and the uncomfortable fact that we already live inside whichever answer we choose, whether we’ve examined it or not.
Frequently Asked Questions
Q: What is the trolley problem and why do most people answer differently in the “fat man” version?
A: The trolley problem asks whether you should divert a runaway trolley to kill one person instead of five. Most people say yes. In the “fat man” variant, you must push an innocent bystander onto the track—most say no. The shift suggests many people care about more than body counts: direct killing, involving nonparticipants, and the felt difference between redirecting harm and using someone as a means.
Q: What is the difference between consequentialist and categorical moral reasoning?
A: Consequentialist reasoning judges acts by their outcomes—e.g., five lives saved outweigh one lost. Categorical reasoning holds that some acts (like killing an innocent person) are wrong regardless of results. Utilitarianism is the main consequentialist framework; Immanuel Kant is the key figure for categorical duties and rights.
Q: What happened in Queen v. Dudley and Stephens?
A: In 1884, after their ship sank, Captain Thomas Dudley and First Mate Edwin Stephens killed and ate 17-year-old cabin boy Richard Parker to survive starvation at sea. They argued necessity; prosecutors called it murder. The case is still debated in law and philosophy classes as a real-world test of utilitarian logic, consent, fair procedure, and whether murder can ever be morally justified to save others.
