The Trolley Problem: Why One of Philosophy’s Strangest Puzzles Won’t Leave You Alone

Introduction

Somewhere in the world right now, a philosophy professor is drawing a stick-figure trolley on a whiteboard, and somewhere in that room a student is quietly deciding that philosophers are ridiculous people. Five figures tied to one track. One figure tied to another. A lever. A runaway vehicle that has apparently never been serviced. It is an absurd scenario, it has never happened, it will never happen, and it is possibly the most productive piece of nonsense ever invented.

Here's what makes it interesting, and it isn't the answer. Almost everyone pulls the lever. Around ninety percent of people, across countries and cultures and levels of education, say they would divert the trolley to kill one person instead of five. Then the scenario changes slightly. Now you're standing on a footbridge above the track, and the only way to stop the trolley is to push a large man off the bridge into its path. Same arithmetic. One dies, five live. And now most people refuse.

That gap is the whole point. You have just discovered that your moral judgment is not a single consistent system but at least two systems that disagree with each other, and you did it in about eleven seconds without any training. The trolley problem isn't a test with a right answer. It's a diagnostic tool, and the thing it diagnoses is you.

This article traces where the puzzle came from, which is not where most people assume, what it has revealed about how human beings actually make moral decisions, why the scenario got taken out of the seminar room and handed to engineers building self-driving cars, and why a growing number of serious philosophers now think the whole enterprise has been a distraction. You will not finish it knowing what to do. You will finish it knowing considerably more about why you don't.

A word about how you're going to read it. Close reading means slowing down deliberately and attending to how a text is built, not merely what it reports. Most language learners hit a wall somewhere around the upper-intermediate level, and the wall is not made of vocabulary. You've read plenty. You understand almost everything. The problem is that you understand it at a single depth, the way you'd read a train timetable, and so you miss the argument being smuggled inside a word choice, the concession that isn't really a concession, the sentence that delays its main verb because delay creates pressure.

Depth is what breaks the plateau. Stopping to ask why the writer used "smuggled" instead of "included." Noticing that a paragraph is structured as a trap. Asking what a question assumes before it asks anything. That attention transfers directly into your speaking and writing, because you cannot deploy a technique you have never consciously seen. The questions after each paragraph are the actual lesson here. Attempt them before you read the analysis. Being wrong on your own is worth far more than being right by peeking.

The Trolley Problem

The first thing worth knowing about the trolley problem is that it was never supposed to be about trolleys, and the woman who invented it would probably be irritated by what happened next. In 1967, the British philosopher Philippa Foot published an essay about the doctrine of double effect, and her actual subject was abortion. She needed a clean example to isolate a specific intuition, so she offered one: the driver of a runaway tram who can steer onto a track where one man will die instead of a track where five will. That was it. A single illustration, a few sentences long, buried in a serious paper about something else entirely. Foot spent the rest of her career on virtue ethics and rarely returned to it. The example escaped, bred in captivity, and now has more descendants than she had publications.

Close reading questions: 

• What is the effect of describing the example as something that "escaped, bred in captivity"? 
• Why does the writer emphasize that the trolley problem's original subject was abortion, and how does that change your sense of what the puzzle is for? 
• What does the phrase "isolate a specific intuition" suggest about the method of analytic philosophy?

To see what Foot was actually after, you have to know about the doctrine of double effect, a piece of medieval moral reasoning usually traced to Thomas Aquinas. The idea is that there's a genuine moral difference between harm you intend and harm you merely foresee. If you steer the tram toward the one man, you don't want him dead. If he miraculously leapt free, you'd be delighted. His death is a side effect of your attempt to save five. But if you push a man in front of a trolley to stop it, his death isn't a side effect at all. It's your instrument. You need him to be hit. If he leapt free, your plan would fail. That distinction sounds like hair-splitting until you notice it's doing enormous work in real life, underwriting the difference between palliative sedation and euthanasia, between a targeted strike and a terror bombing, between letting a patient go and killing them.

Close reading questions: 

• How does the "leapt free" test work as an argumentative device, and why is it more persuasive than an abstract definition of intention? 
• The paragraph moves from "hair-splitting" to three real-world applications: what is that structural move doing to the reader? 
• Can you construct a case where the intention/foresight distinction seems to produce the wrong answer?

Then Judith Jarvis Thomson got hold of it, and things got weird in the most productive possible way. In papers published in 1976 and 1985, she generated variations the way a chemist generates compounds, changing one variable at a time to see what our intuitions were actually tracking. The footbridge version, where you push a heavy man to his death, is hers. So is the loop track, where the trolley diverts onto a side rail that curves back to the main line, and the single man's body is the only thing that will stop it, which means you are using him as a tool while merely turning a wheel. So is the surgeon case: a doctor with five dying patients and one healthy man in the waiting room whose organs would save them all. Nobody defends the surgeon. Almost everybody pulls the lever. The arithmetic is identical in all of them, which means the arithmetic is not what we are responding to.

Close reading questions: 

• Why does the writer compare Thomson's method to chemistry, and what does that comparison claim about philosophy as a discipline? 
• What is the loop track variation designed to prove, and against whom is it aimed? 
• If "the arithmetic is not what we are responding to," what are the candidate answers, and how would you test between them?

The obvious explanation is that our intuitions are simply confused, and that we should trust the math and get on with it. This is roughly the utilitarian position, and it deserves better than the caricature it usually gets. Utilitarianism says that what makes an action right is its consequences, specifically how much wellbeing it produces, and that everyone's wellbeing counts equally. That last part was radical when Bentham wrote it and remains uncomfortable now, since it means your child's welfare counts exactly as much as a stranger's child, and no more. Against this stands the broadly Kantian tradition, which holds that people are not quantities to be summed. There is something you owe a person simply as a person, and it does not disappear because four other people would benefit. On this view, pushing the man off the bridge doesn't just produce a bad outcome. It commits a specific wrong against him, and the fact that five strangers benefit is, morally speaking, none of his business.

Close reading questions: 

• What is the writer doing by insisting that utilitarianism "deserves better than the caricature it usually gets," and how does that affect the paragraph's credibility? 
• Explain the phrase "none of his business" in your own words: what moral claim is compressed into that idiom? 
• Which of these two positions does the paragraph's language subtly favor, and what specific words give it away?

In 2001, a neuroscientist named Joshua Greene put people in fMRI scanners and ran them through both versions, and the results reshaped the entire conversation. The switch case, the impersonal one, lit up regions associated with abstract reasoning and calculation. The footbridge case lit up regions associated with emotional processing and social judgment. Greene's interpretation, which he called the dual-process theory, is that we are running two moral systems on the same hardware. One is fast, ancient, emotional, and reacts violently to up-close personal violence, because our ancestors spent a very long time in small groups where shoving someone was a live option and pulling a lever was not. The other is slow, abstract, and capable of arithmetic. They frequently disagree, and what we experience as moral conviction is often just one of them shouting louder.

Close reading questions: 

• What does the phrase "running two moral systems on the same hardware" borrow from, and what does that metaphor gain and lose?
• How does the evolutionary explanation function in the argument, and is it doing descriptive work or normative work? 
• What would count as evidence against the dual-process theory?

Here is where it gets genuinely uncomfortable. Greene draws a conclusion many people find hard to accept: if our resistance to pushing the man is produced by an emotional module that evolved for a world without levers, then that resistance is not evidence about morality. It's evidence about our evolutionary history. The intuition is real, but it isn't tracking anything true. On this view, the deontologist who insists that pushing is simply wrong is doing sophisticated post-hoc rationalization of a gut reaction, dressing up a reflex in philosophical clothing. That is a serious charge. It is also, if you think about it for a moment, a charge that cuts both ways, because the capacity for cold arithmetic evolved too, and nobody has explained why a system that computes should be more truth-tracking than a system that feels.

Close reading questions: 

• What is the function of the sentence "That is a serious charge" as a standalone unit, and how does the paragraph use it as a hinge? 
• Reconstruct the counterargument in the final sentence as a formal objection: what exactly is being claimed? 
• Why might a writer choose "dressing up a reflex in philosophical clothing" instead of a technical term like "confabulation"?

Meanwhile, something strange happened to the trolley problem outside philosophy. Around 2015, as self-driving cars stopped being science fiction, journalists discovered the scenario and would not let go of it. Suddenly every article about autonomous vehicles asked whether the car should swerve to kill one pedestrian rather than five, and MIT built the Moral Machine, an online platform that collected around forty million decisions from millions of people in over two hundred countries. The results were fascinating and slightly alarming. Preferences varied by region in patterned ways, with some clusters strongly prioritizing the young over the old and others showing the opposite tendency, and nearly everywhere people preferred to spare the lawful over the jaywalking. Engineers, for their part, mostly rolled their eyes. Their actual problem is not choosing between victims. It's detecting a cyclist in the rain.

Close reading questions: 

• Why does the writer end this paragraph with the engineers' complaint rather than the survey results, and what is the effect of that ordering? 
• What does the phrase "would not let go of it" imply about media attention, and is that implication fair? 
• If moral preferences vary systematically by region, what follows for the design of a global product, and what does not follow?

The engineers have a point, and it opens onto the strongest criticism of the whole enterprise. Trolley cases are built by stripping away everything that makes real moral situations difficult. You know the outcomes with certainty. You have no relationship with anyone involved. There is no time to consult, no possibility of a third option, no aftermath, no institution, no history. Real moral life is almost entirely composed of the things the thought experiment deletes: uncertainty, relationships, the option of asking someone, the fact that you have to keep living with the people involved afterward. Critics argue that training your moral reasoning on trolley cases is like training for a marathon on a treadmill in a room with no weather, no hills, and no other runners. You will get very good at something, and it may not be the thing you needed.

Close reading questions: 

• How does the treadmill analogy work, and where does it break down? 
• The paragraph lists what thought experiments remove: which item on that list do you consider the most damaging omission, and why? 
• Is there a defense of abstraction available here that the paragraph has not offered?

But that criticism, powerful as it is, may prove less than it seems. The trolley problem was never meant to be a simulator. It's an instrument, and instruments are supposed to be artificial. A physicist studying gravity uses a vacuum chamber precisely because air resistance is a real feature of the world that obscures the thing being measured. Foot removed relationships and uncertainty for the same reason a chemist removes contaminants: not because contaminants don't exist, but because you cannot see what you're testing while they're present. The question isn't whether the scenario is realistic. It's whether what you learn in the vacuum still holds when you let the air back in. Sometimes it does. Sometimes it spectacularly doesn't, and finding out which is which is most of what applied ethics consists of.

Close reading questions: 

• How does this paragraph use the vacuum chamber analogy to answer the treadmill analogy, and which is more apt? 
• What does "may prove less than it seems" concede and what does it withhold? 
• The final sentence claims that distinguishing the two cases is "most of what applied ethics consists of": is that claim defended anywhere, and does it need to be?

So what is the trolley problem actually good for, if it can't tell you what to do? Something more useful, probably. It reveals that you are not one moral reasoner but several, in uneasy coalition. It shows you that you hold principles you have never articulated and could not have predicted, since almost nobody guesses in advance that they will treat pushing and pulling differently. It demonstrates that the difference between what you do and what you allow, between what you intend and what you foresee, between touching someone and operating a machine that touches them, matters enormously to you, whether or not you can justify it. And it does all of that in under a minute, using a vehicle that doesn't exist and a scenario that never happens, which is a remarkable return on an absurd investment.

Close reading questions: 

• What is the rhetorical effect of the list of three "it shows/reveals/demonstrates" clauses, and why does the writer place the shortest claim first? 
• The phrase "uneasy coalition" is doing precise work: what does a coalition imply that a "system" would not? 
• Do you accept that a thought experiment can teach you something about yourself that you could not have learned by introspection?

Here's the thing nobody tells you about the trolley problem. In every version, you are standing beside the track, uninvolved, when the situation arrives. You did not tie anyone to the rails. You did not fail to maintain the brakes. You are a bystander granted sudden terrible power over strangers, and the puzzle asks only what you should do in the next four seconds. But almost no real moral catastrophe works that way. The people who end up at the lever usually got there through a thousand small ordinary decisions made years earlier, by themselves and by others, none of which felt like a moral crisis at the time. So the question worth carrying out of here isn't which way you'd pull. It's this: which of the levers you are standing next to right now have you not noticed, because nobody has drawn you a diagram of them?

The Deep Dive Analysis

What is the effect of describing the example as something that "escaped, bred in captivity"?

The metaphor treats the thought experiment as a laboratory organism, which does several things at once. It implies that Foot created something in a controlled setting for a limited purpose, that it got loose, and that it then reproduced without supervision, which is a compact history of the trolley problem's actual reception. "Bred in captivity" is the sharper half of the phrase because it contains a small paradox: the escape was into philosophy departments, which are themselves a kind of enclosure, so the variations proliferated in an environment as artificial as the original. The metaphor also carries a note of faint alarm, the way we talk about invasive species, which prepares the reader for the article's later suggestion that the puzzle may have crowded out more useful work. Notice the grammatical economy. Three verbs in a row, "escaped, bred, and now has," compress fifty years into a single sentence, and the shift from past tense into present is what makes the history feel like it's still happening.

Why does the writer emphasize that the original subject was abortion, and how does that change what the puzzle is for?

Because it reframes the entire enterprise. Read as a freestanding brain teaser, the trolley problem looks like recreational philosophy. Read as an instrument built to test a principle that governs abortion, end-of-life care, and the ethics of warfare, it becomes obvious that the stakes were always concrete and the abstraction was a means, not an end. The emphasis also functions as a quiet rebuke to the puzzle's popular reception, since the version most people encounter has been stripped of the very context that justified it. There is a broader lesson for readers about intellectual history: ideas routinely detach from their originating problems and acquire lives that would surprise their authors, and knowing the original problem often reveals what the idea was actually shaped to do. When you encounter a famous concept, it's worth asking what question it was first an answer to.

What does "isolate a specific intuition" suggest about the method of analytic philosophy?

It presents philosophy as something closer to experimental science than to contemplation. To isolate is to separate a variable from its surroundings so it can be observed independently, which is the language of the laboratory. The implied method is analytic in the literal sense: break a complex moral judgment into components, vary one component, and observe whether the judgment changes. This is genuinely how much twentieth-century Anglophone philosophy operated, and it explains the discipline's fondness for bizarre hypotheticals, which are not eccentricity but instrumentation. The phrase also quietly sets up the article's later vacuum chamber defense, so a careful reader will notice that the argument's conclusion is seeded in its second paragraph. Good expository writing frequently does this, planting a frame early so that a later argument feels like a recognition rather than an assertion.

How does the "leapt free" test work as an argumentative device, and why is it more persuasive than an abstract definition of intention?

It converts a metaphysical question into a practical one. Defining intention properly is notoriously difficult, but the counterfactual test sidesteps the definitional swamp entirely: imagine the victim escaping, and ask whether your plan still works. In the switch case, you'd be relieved. In the footbridge case, you'd have to find another heavy man. The difference is immediately felt rather than argued, which is what gives the test its force. Philosophers know it as a version of the counterfactual test for intention, and it is not without problems, as the loop track case is specifically designed to show. As a piece of writing, though, it demonstrates a transferable technique: when you need to make an abstract distinction vivid, don't define it, construct a scenario in which the two sides come apart and let the reader feel the difference. Note the tense work as well, with "if he miraculously leapt free, you'd be delighted" using the second conditional to build a hypothetical the reader inhabits rather than evaluates.

The paragraph moves from "hair-splitting" to three real-world applications. What is that structural move doing?

It is a preemptive concession, one of the most useful moves in persuasive English. The writer voices the reader's likely objection, that this is precious academic distinction-drawing, and then immediately demonstrates that the distinction underwrites decisions in medicine and warfare where people actually die. By articulating the objection first, the writer removes the reader's ability to hold it privately as a reason to disengage. The three examples are also carefully ordered, moving from medicine, which is intimate and familiar, to warfare, which is institutional and remote, and ending on the starkest phrasing. The rhythm of a tricolon, three parallel items, creates a sense of completeness and closure that two would not. When you're arguing in writing, look for the sentence where your reader is most likely to check out, and put their objection in your own mouth just before it.

Can you construct a case where the intention/foresight distinction produces the wrong answer?

Several are standard in the literature. The most damaging is the terror bomber and tactical bomber pair: one bombs a munitions factory knowing civilians nearby will die, the other bombs civilians deliberately to break morale. Double effect says the first is permissible and the second is not, even if the tactical bomber kills far more people. Many find this unbearable, because it appears to let the size of the harm be trumped by the bomber's mental state. A second case: a doctor who withholds treatment intending the patient's death but who could describe the same act as merely foreseeing it, which suggests the doctrine may be gameable by clever redescription of one's own intentions. This is sometimes called the closeness problem. The honest position is that the distinction captures something most people genuinely believe while resisting precise formulation, which is a familiar situation in ethics and a reason to hold it loosely rather than either adopting or discarding it wholesale.

Why compare Thomson's method to chemistry, and what does that claim about philosophy?

The comparison claims systematicity. A chemist doesn't invent compounds at random; she varies one element to determine what a property depends on. Saying Thomson worked this way asserts that the proliferation of trolley variations, which looks from outside like philosophers amusing themselves, was actually a controlled search for the variable driving moral judgment. The metaphor also carries a subtle defense of the discipline against the charge of frivolity, which the article will address directly later. What the comparison loses is worth naming: chemistry has an external check, since the compound either forms or it doesn't, whereas the trolley method's data is intuition, which is exactly what's under suspicion. The circularity there is real, and a fair-minded reader should hold the analogy at arm's length even while accepting its main point.

What is the loop track variation designed to prove, and against whom is it aimed?

It is aimed at defenders of the doctrine of double effect, and its purpose is to collapse the distinction the doctrine depends on. In the loop case, you pull a lever, exactly as in the standard version, but the track curves back, so the trolley will still hit the five unless the one man's body stops it. You are therefore using him as a means, precisely as in the footbridge case, while performing the physically impersonal action of the switch case. If your intuition says pulling is still permissible here, then the means/side-effect distinction cannot be what's driving your judgment, and something else is, most likely the physical directness of the act. It's an elegant piece of philosophical engineering: rather than arguing that a distinction is wrong, Thomson constructs a case where it comes apart from the intuition it was supposed to explain. That is the more devastating form of objection, and it's worth learning to recognize.

If "the arithmetic is not what we are responding to," what are the candidate answers, and how would you test between them?

The leading candidates are physical contact, whether you touch the victim; personal force, whether the harm comes from your body rather than a mechanism; the means/side-effect distinction; spatial proximity; and the presence of an intervening agent or object. Testing between them requires constructing cases that separate the variables, which is exactly the trolley industry's business. Greene and colleagues did this empirically, adding a version where you push the man with a pole, which preserves personal force while removing direct contact, and a trapdoor version where you open a hatch beneath him, which removes personal force while preserving the use of him as a means. Judgments tracked personal force more closely than the means/end distinction, which is bad news for double effect as a description of what people actually think, though not necessarily as a claim about what they should think. Note the distinction being drawn there, between describing moral psychology and justifying moral principles, because conflating the two is one of the commonest errors in this whole conversation.

What is the writer doing by insisting utilitarianism "deserves better than the caricature it usually gets"?

Establishing fairness as a credential. In any piece that will eventually take a position, treating the opposing view generously early on buys credibility for later criticism, because the reader has evidence that the writer isn't merely scoring points. It also does substantive work by warning against a specific misreading, since utilitarianism is frequently portrayed as cold, calculating, and indifferent to persons, when historically it was a radical egalitarian doctrine that extended moral consideration to the poor, to prisoners, to animals, and to women at a time when this was genuinely unpopular. The rhetorical structure to notice is concede-then-complicate. Watch how the paragraph then presents the Kantian objection without dismissing the utilitarian one, leaving the reader in genuine tension rather than delivering a verdict.

Explain "none of his business" in your own words: what moral claim is compressed into that idiom?

The idiom normally means something is not your concern, and applying it here produces a small jolt, because the man on the bridge is very much concerned in the ordinary sense. He's about to die. What's being claimed is deeper: that from his standpoint, the benefit accruing to five strangers cannot serve as a justification addressed to him. He is owed a reason, and "other people will be better off" is not a reason he could reasonably be expected to accept. This is close to the contractualist idea, associated with T. M. Scanlon, that an act is wrong if it could be reasonably rejected by anyone affected. Compressing that into a colloquial idiom is a deliberate stylistic choice, making an abstract principle land with the force of a shrug. It also demonstrates something about advanced English: idioms used slightly against their normal sense produce meaning that literal phrasing cannot, and recognizing that misuse is deliberate rather than mistaken is a genuine reading skill.

Which position does the paragraph subtly favor, and what words give it away?

Slightly the Kantian one, and the tell is in the verbs and nouns. Utilitarianism gets "counts," "produces," "summed," a vocabulary of accounting. The Kantian view gets "owed," "commits a specific wrong against him," a vocabulary of relationship and obligation. The phrase "quantities to be summed" is the pivot, since it characterizes the utilitarian view in terms its own adherents would resist. The paragraph also gives the Kantian position the final word, and final position is emphatic position in English prose. This is worth practicing as a reading habit: when a writer presents two views apparently even-handedly, examine the verbs assigned to each and note which view is allowed to close the paragraph. Bias in careful writing rarely appears in the claims. It lives in the diction and the running order.

What does "running two moral systems on the same hardware" borrow from, and what does the metaphor gain and lose?

It borrows from computing, treating the brain as hardware and cognitive processes as software. The gain is enormous clarity: the metaphor instantly explains how a single person can hold contradictory judgments without being irrational or dishonest, since two programs can run simultaneously and produce different outputs. It also makes the phenomenon feel mechanical rather than shameful, which lowers the reader's defenses. The loss is that the metaphor implies a separability that may not exist, as emotion and reasoning are not cleanly modular in the brain, and dual-process theory has been substantially complicated since its popular heyday. It also smuggles in the idea that one system might be a "bug," which is precisely the contested normative question. Computing metaphors are so pervasive in modern English that they slip past unexamined, and it's worth noticing every time you meet one what picture of the mind it's quietly installing.

How does the evolutionary explanation function, and is it descriptive or normative?

In the article it starts descriptive and turns normative, which is the crucial move to catch. Describing why we have an aversion to close-up violence is a historical claim about our ancestors. Concluding that the aversion is therefore not evidence about morality is a philosophical claim that requires an additional premise, roughly that a judgment produced by a process not aimed at truth is unjustified. That premise is contestable and is exactly where the debate lives. The general hazard here is the naturalistic fallacy and its inverse, the genetic fallacy: explaining the origin of a belief neither justifies nor refutes it. When you meet an evolutionary explanation of a human tendency, whether in an article about morality, gender, or economics, the question to ask is whether the writer has slid from "this is why we feel it" to "therefore the feeling is worthless," and whether the extra premise required for that slide has been defended.

What would count as evidence against the dual-process theory?

Several kinds. Behaviorally, if people under time pressure or cognitive load did not become more likely to give the emotional answer, a central prediction would fail. Neurologically, if patients with damage to emotional processing regions did not show elevated utilitarian judgments, that would undercut it, though in fact such findings were part of the theory's original support. Methodologically, the strongest challenges have come from failures to replicate specific findings, from the observation that the footbridge and switch cases differ in many respects beyond emotional salience, and from arguments that the so-called utilitarian responders are not being more rational at all but scoring higher on measures of reduced empathy. Asking what would falsify a theory is the single most useful habit in critical reading, because it separates claims that are doing scientific work from claims that merely sound scientific.

What is the function of "That is a serious charge" as a standalone sentence, and how does the paragraph use it as a hinge?

It performs three functions in six words. It marks a change of direction, signaling that the paragraph is about to turn against the argument it has just built. It grants weight to that argument before criticizing it, which is a fairness move. And it slows the reader down through sheer brevity, since a short declarative sentence after several long ones creates a pause with physical force. This is a structural device worth stealing: the hinge sentence, deliberately plain, placed at the pivot point of a paragraph. Note also the demonstrative pronoun "that," which points backward and gathers up the preceding sentences into a single referent. Skilled English writers use "this" and "that" as sentence openers precisely for this bundling function, and doing it well requires that the referent be unambiguous, which is why unclear "this" is one of the most common weaknesses in intermediate writing.

Reconstruct the counterargument in the final sentence as a formal objection.

The objection runs like this. Greene's debunking argument requires that emotional moral responses are unreliable because they were shaped by evolutionary pressures unrelated to moral truth. But the capacity for abstract cost-benefit reasoning was also shaped by evolutionary pressures unrelated to moral truth, since it presumably evolved for foraging, tool use, and social coordination rather than for identifying moral facts. Therefore, if evolutionary origin undermines a faculty's authority, it undermines both faculties equally, and Greene's argument proves too much. To escape this, the utilitarian needs an independent reason why calculation is more truth-tracking than emotion in the moral domain, and it is not obvious what that reason would be without assuming utilitarianism at the outset. This is a real objection in the literature, and Greene has responses to it, chiefly that we should trust the general-purpose reasoning system more precisely when facing novel problems the specialized system wasn't built for. The point for a reader is that the paragraph's closing sentence is not a rhetorical flourish. It's a compressed argument, and you should be able to unpack it.

Why "dressing up a reflex in philosophical clothing" instead of "confabulation"?

Register and reach. "Confabulation" is precise and would satisfy a specialist, but it requires the reader to already know it, and it lacks any picture. The clothing metaphor makes the accusation visible, since you can see the reflex underneath and the philosophy draped over it, and it carries an implication of pretension that the technical term does not. The choice also matches the article's voice throughout, which favors vivid ordinary language over jargon while still making technical points. Note that the metaphor's condescension is deliberate, because the article is characterizing an accusation, not endorsing it, and the slightly unfair phrasing prepares the reader to sympathize with the rebuttal that immediately follows. That's a subtle piece of engineering: the writer states the opposing charge in language just harsh enough to make the reader want it answered.

Why end the self-driving car paragraph with the engineers' complaint rather than the survey results?

Because final position is where emphasis lives, and the writer wants the deflation, not the data, to be what the reader carries forward. The Moral Machine results are genuinely interesting, but ending there would suggest the trolley problem had been vindicated as an engineering specification. Ending with "detecting a cyclist in the rain" punctures that, and it does so with a concrete image after a paragraph of large abstract numbers, which is why it lands. The sentence is also short and unadorned following longer ones, the same rhythmic device used earlier. There's a deeper argumentative purpose too, since this deflation opens the door for the next paragraph's criticism of thought experiments generally. Paragraph endings in well-constructed prose are almost never neutral; they are handoffs, and tracking what each one sets up is a reliable way to see a piece's architecture.

What does "would not let go of it" imply about media attention, and is that fair?

It implies obsessive, slightly undignified persistence, like a dog with a stick, and the implication is only partly fair. The media latched onto the trolley problem because it's visual, quickly explained, morally alarming, and generates engagement, which is a legitimate description of how attention economies work. But it's also true that journalists were responding to genuine public unease about delegating life-and-death decisions to machines, and the trolley problem was the closest available vocabulary for that unease. So the phrase is a fair criticism of the coverage's proportions and an unfair dismissal of its motive. A reader should notice that the article never actually argues for the criticism it insinuates through this idiom, which is a normal and mostly harmless feature of essayistic prose, but worth catching. Insinuation through connotation is how a great deal of persuasion happens without any claim being made that could be challenged.

If moral preferences vary systematically by region, what follows for a global product, and what does not follow?

What follows is practical difficulty. A company deploying identical software worldwide will impose one set of priorities on populations with different intuitions, and this is a genuine political problem about legitimacy and consent, not merely a technical one. What does not follow is moral relativism. That people disagree about something establishes nothing about whether there's a fact of the matter, since populations have disagreed about slavery, about the status of women, and about whether the sun moves. The descriptive finding and the normative conclusion are separate, and the slide between them is one of the most common errors in popular writing about moral psychology. What might follow, more modestly, is a procedural conclusion: where reasonable disagreement is deep and persistent, decisions should perhaps be made through democratic processes rather than by engineers or ethicists, which is an argument about who decides rather than about what's true.

How does the treadmill analogy work, and where does it break down?

It works by mapping controlled training onto controlled reasoning, with the shared implication that removing variables makes you skilled at an artificial version of the task. The specifics are well chosen, since weather, hills, and other runners correspond neatly to uncertainty, difficulty, and other people, and the payoff line, "you will get very good at something, and it may not be the thing you needed," is precise about the danger without overstating it. It breaks down in two places. First, treadmill training genuinely does build cardiovascular fitness that transfers to real running, which actually undercuts the criticism, and a hostile reader can turn the analogy against the paragraph. Second, and more fundamentally, the trolley problem was never intended as training. It's a measuring device, and measuring devices are not supposed to resemble the world, which is exactly the objection the next paragraph makes. Analogies are the most persuasive and most dangerous device in expository writing, and the discipline of asking where one breaks down should be automatic.

Which deleted feature is the most damaging omission?

A defensible answer is uncertainty, because it is not one feature among many but the medium in which real moral decisions occur. In trolley cases you know that five will die and one will die. In life you almost never know, and most of the hardest moral questions are questions about how much risk you may impose on others, how confident you must be before acting, and who bears the cost of your being wrong. An ethics built entirely on certain outcomes has nothing to say about the ordinary situation. A strong rival answer is the absence of relationships, since much of moral life concerns special obligations to particular people, and the trolley's anonymous victims make partiality invisible. A third is the absence of aftermath: in the thought experiment the story ends at the lever, whereas in life you live with what you did, and the moral significance of guilt, repair, and apology is entirely outside the frame. Being able to argue for more than one answer here, and to rank them with reasons, is precisely the analytical skill this exercise is meant to build.

Is there a defense of abstraction the paragraph has not offered?

Yes, and the article supplies part of it in the next paragraph but not all. The unstated defense is pedagogical rather than epistemic: thought experiments are effective because they are memorable and portable, and a principle you can carry in your head as an image will influence your behavior more than a principle you understood once in a nuanced discussion and then forgot. There's also a defense from disagreement management, since abstract cases let people with radically different worldviews locate exactly where they diverge, which is much harder in a real case cluttered with contested facts. And there's a defense from moral safety: you can explore an intuition about killing without anyone being harmed, which is not true of empirical ethics. Noticing which available arguments a text has not made is an advanced reading skill, and it's how you move from following an argument to evaluating one.

How does the vacuum chamber analogy answer the treadmill analogy, and which is more apt?

It answers by reclassifying the object. The treadmill analogy assumes the trolley problem is training, where realism matters, and the vacuum chamber assumes it's measurement, where artificiality is the point. Since the two analogies disagree about what kind of thing the thought experiment is, the debate isn't really about realism at all. It's about the purpose of moral philosophy, which is the real disagreement wearing a disguise. The vacuum chamber is more apt to Foot's original intent, since she was isolating a principle, and the treadmill is more apt to the trolley problem's modern classroom use, where it genuinely is presented as moral training. Both are therefore right about different things, which is why the article concedes rather than dismisses. When two analogies collide in an argument, the productive move is not to pick a winner but to ask what each one assumes about the nature of the thing being described.

What does "may prove less than it seems" concede and what does it withhold?

It concedes that the criticism has force, signaled by the modal "may" and the earlier "powerful as it is," while withholding agreement that the criticism is decisive. The construction is a hedge, and hedging in academic and essayistic English is a technical skill rather than a weakness, because overclaiming invites easy refutation while a well-placed modal keeps a claim defensible. Note the specific verb: "prove" belongs to the vocabulary of logic and evidence, so the sentence is saying the objection may not establish what it appears to establish, which is a claim about argumentative reach rather than about truth. The criticism might still be correct and simply not proven by this route. That is a fine distinction, and picking up on it is exactly the kind of reading that separates C1 from C2. Learners who want to sound advanced in writing should study modal verbs and hedging constructions with the seriousness usually reserved for vocabulary lists.

Is the claim about applied ethics defended, and does it need to be?

It isn't defended, and it doesn't strictly need to be, because it functions as a closing characterization rather than a load-bearing premise. The paragraph's argument is complete without it, and the sentence serves to place the preceding point in a wider professional context, telling the reader that this is what the field actually does. That said, an alert reader should register that a substantial claim about an entire discipline has just been asserted in passing. Essayistic writing routinely makes such gestures, and they are usually acceptable when they summarize rather than establish. The test is whether removing the sentence would damage the argument, and here it would not. When you are evaluating any piece of nonfiction, distinguishing load-bearing claims from decorative ones tells you where to concentrate your skepticism, and it also tells you, when you write, where you can afford a flourish and where you cannot.

What is the rhetorical effect of the three parallel clauses, and why is the shortest claim placed first?

The parallel structure creates accumulation, each clause reinforcing the pattern so that the three land as a single unified case rather than three separate assertions. The verbs escalate deliberately, from "reveals" to "shows" to "demonstrates," with the last being the strongest evidential term, so the sequence gains confidence as it proceeds. Placing the shortest first and the longest last follows the principle of end weight, which governs a great deal of English prose rhythm: heavier, more complex elements belong at the end of a series, and reversing that order produces a sense of anticlimax. The technique is easy to overuse, and three is generally the maximum before parallelism starts to sound like a political speech. If you want to internalize this, take any list you've written and try reordering it purely by length, then read both aloud.

What does "uneasy coalition" imply that "system" would not?

A system implies design, integration, and a single purpose, with parts that were built to work together. A coalition implies separate parties with distinct interests who have joined for practical reasons and may break apart. "Uneasy" adds that the alliance is strained and possibly temporary. The phrase therefore claims that your moral faculties were not designed as a unit, which is exactly what the evolutionary story suggests, and that their cooperation is a negotiated truce rather than a functioning machine. It also carries a political flavor, implying internal argument, lobbying, and shifting majorities, which is a surprisingly good description of what deliberation feels like from the inside. This is metaphor doing genuine analytical work rather than ornamenting an already-complete idea, and it's the standard worth holding your own metaphors to.

Can a thought experiment teach you something about yourself that introspection could not?

The strong case says yes, and the trolley problem is the evidence. Almost nobody, asked in advance to describe their moral principles, predicts that they will treat pulling and pushing differently. The principle is real, operative, and invisible until a scenario is constructed that forces it into the open, which is precisely what introspection cannot do, since introspection can only report what you already have access to. This connects to a large body of psychological work suggesting that people are poor witnesses to the causes of their own judgments and frequently generate plausible explanations after the fact. The skeptical case is worth stating too: the thought experiment may not reveal a pre-existing principle so much as create a response through its framing, and asking people about bizarre hypotheticals may tell you mainly how they answer bizarre hypotheticals. The reasonable position is that thought experiments are real evidence about moral psychology, but evidence of a specific and limited kind, and treating them as a window onto the soul overstates what they can do.

Writing Challenge

Write between 600 and 800 words constructing an original thought experiment and then arguing about what it reveals.

Choose a moral question you actually find difficult, not one where you already know your answer. Build a scenario that isolates a single variable, in the way Foot isolated intention and Thomson isolated the use of a person as a means. Then produce a second version of your scenario in which you change exactly one thing, and show that the change flips or weakens the intuition. That pair is your instrument, and the essay is the report of what it measured.

Your piece must do four things. It should state clearly what the two versions hold constant and what they vary, since a thought experiment that changes several things at once measures nothing. It should include at least one preemptive concession, in which you voice the strongest objection to your scenario before your reader can. It should include one carefully hedged claim using modal verbs, and you should be able to say precisely what the hedge concedes and what it withholds. And it should end by drawing a distinction between what your experiment shows about moral psychology and what, if anything, it shows about morality.

Two constraints on style. Somewhere in the piece, use an idiom slightly against its ordinary sense to make an abstract point land, as "none of his business" was used above. And include one hinge sentence of six words or fewer at the exact point where your argument changes direction.

Speaking Challenge

Prepare and deliver a five-to-six-minute spoken argument, recorded if you can, on this proposition: "An intuition you cannot justify is still a reason."

You may argue for or against, but you must do four things regardless of your position. Open by defining what you mean by an intuition, distinguishing it from a preference, a habit, and a prejudice, because the whole argument turns on whether those are different things. Use at least one example from outside the trolley problem, ideally from law, medicine, family life, or your own culture's moral vocabulary. Spend a full minute steelmanning the position you reject, presenting it in its strongest form rather than its most convenient one. And explicitly address the debunking argument: if you can show that a moral intuition was produced by evolution, upbringing, or social conditioning, have you shown that it's worthless, and if not, why not?

For the last minute, set the argument aside and answer this directly. Describe a moral judgment you hold that you have never been able to justify to someone who disagrees with you. Say what happens in you when you try. Then say honestly whether you think you've been reasoning all along or defending something that arrived before the reasoning did, and whether it would change anything if you knew.