Ethics
Trolley Problem
Thought experiment about ethical dilemmas — utilitarian vs deontological intuitions
The trolley problem is a famous ethical thought experiment introduced by Philippa Foot (1967), refined by Judith Jarvis Thomson. Original: trolley about to kill 5 workers; you can divert it to side track, killing 1 instead. Most say: divert. Variation: footbridge — you can push large person off bridge to stop trolley (saves 5; kills 1). Most refuse. Same numbers, very different reactions. Reveals: tension between consequentialism (always 5 > 1) and deontological intuitions (active killing different from letting die). Used to: clarify ethical theories, study moral psychology, debate AI ethics.
- OriginatorPhilippa Foot (1967)
- Refined byJudith Jarvis Thomson
- Original variantSwitch trolley (most agree)
- Footbridge variantPush person (most refuse)
- DemonstratesTension between consequentialism and deontology
- UseMoral psychology, AI ethics
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
Why trolley problem matters
- Moral psychology. Studying intuitions empirically.
- Ethics teaching. Iconic thought experiment.
- AI ethics. Programming moral decisions.
- Self-driving cars. Real-world version.
- Theory clarification. Tests utilitarian vs deontological.
- Public engagement. Accessible philosophy.
- Cross-cultural psychology. Universal intuitions?
Common misconceptions
- One right answer. Designed to reveal intuitions.
- Settles ethics. Test for theories, not solution.
- Just numbers. Many factors matter to people.
- Practical guide. Artificial; real ethics messier.
- Foot intended specific answer. Used for various purposes.
- Apply to all autonomous decisions. Real cases include uncertainty.
Frequently asked questions
What's the basic trolley problem?
Trolley out of control on track. Five workers ahead, can't escape. Lever next to you diverts trolley to side track where one worker is. Pull lever or not? If you do: 1 dies. If not: 5 die. Most people: pull lever (5 > 1). Seems straightforward.
What's the footbridge variant?
You're on bridge over track. Trolley approaches; below: 5 workers ahead. Large person standing next to you. Pushing them off would stop trolley (5 saved, 1 dies). Most people refuse. Yet: same numbers as switch case (1 dies, 5 saved). Why different reaction? Reveals: different moral intuitions about active killing vs redirecting harm.
Why do reactions differ?
Multiple explanations. (1) Doctrine of double effect: redirecting harm permitted; using person as means to stop harm prohibited. (2) Killing vs letting die: active causation feels different. (3) Personal force: pushing more visceral than pulling lever. (4) Persons-as-means: footbridge uses person's body as instrument. (5) Naturalistic: evolved repugnance to direct violence. All shed light on moral psychology.
What's the doctrine of double effect?
Roman Catholic moral principle. Action with both good and bad effect: permissible if (1) act itself good or neutral, (2) good effect intended, (3) bad effect not intended (foreseen but not aimed at), (4) good and bad effects proportional. Trolley switch: redirecting good intent; one death foreseen but not intended. Footbridge: using person's death as means — bad intent.
How is moral psychology studied via this?
Joshua Greene used fMRI: footbridge engages emotional brain regions more; switch engages cognitive. Trolley problem shows: emotion and reason interact in moral decisions. Cross-cultural studies: similar reactions globally. Suggests universal moral intuitions; specific framings affect responses. Used: moral psychology research.
How does it apply to AI?
Self-driving cars: must make crash decisions. Save 5 pedestrians at cost of 1 passenger? Mostly utilitarian view if asked abstractly. But many wouldn't buy car that would kill them. Trolley dilemmas in coded ethics. Real world: situations rarely so clean. Still: helps clarify what we want from automated systems.
Is the trolley problem useful?
Mixed. Pro: reveals tensions in ethics; productive thought experiment. Con: too artificial; real ethics involve uncertainty, relationships, longer time horizons. Some philosophers (Bernard Williams): such constructions miss what's important about morality. Many useful insights nonetheless. Probably overused: real moral problems are messier than trolleys.