DEEP READING · READ 46

On Liberty

On Liberty · John Stuart Mill · 1859

中文 →

In One Sentence

This short book answers exactly one question: under what conditions may a society interfere with an individual? Mill's answer leaves almost no room — only to prevent harm to others; "for his own good" is never a sufficient warrant. And the enemy he has in mind is not a tyrant. In a country that already has elections and a parliament, Mill argues, the hardest chains are the ones the majority applies with a clear conscience: opinion, custom, and the flat weight of what everyone thinks.

Where It Sits

Mill (1806–1873) inherited utilitarianism (the doctrine that acts are right insofar as they produce happiness) and then quietly rebuilt it. Greek at three, Latin at eight, raised by his father James Mill on Bentham's blueprint as an instrument of reform, he broke down at twenty: he asked himself whether the total success of every cause he was working for would make him happy, and something in him answered, distinctly, no (his Autobiography, paraphrased). What pulled him out was not a better argument but Wordsworth's poetry — and for the rest of his life he was supplying utilitarianism with the thing it had been missing: the inward growth of a person. On Liberty appeared in 1859, dedicated to his wife Harriet Taylor Mill, who had died the year before and whom he called its joint author. It is liberalism's clearest statement of itself — argued not from natural rights but from what kind of society lets a human being become one.

The Central Claims

• The harm principle: the only purpose for which power can rightfully be exercised over any member of a civilised community, against his will, is to prevent harm to others. His own good, physical or moral, is not a sufficient warrant. You may reason with him, plead with him, avoid him — you may not compel him or punish him. In Mill's phrase, over himself, over his own body and mind, the individual is sovereign (paraphrased).
• Freedom of thought and discussion must be near-absolute — not because opinions are harmless, but because silencing one is a theft from the whole human race. Silence a true opinion and you lose the chance to correct yourself; silence a false one and you lose just as much, because a truth that is never struck against error stops being a live belief and becomes a slogan nobody can justify.
• Individuality is not an eccentricity to be tolerated; it is one of the elements of well-being. A person who merely follows custom has made no choice at all, and a society composed of such people loses the capacity to improve.
• The dangerous oppressor in a modern society is not the state but society itself. Once the tyranny of the majority moves out of law and into opinion and custom, it becomes far harder to escape: it does not execute you, it seeps into the details of daily life until it enslaves the soul itself (paraphrased).

The Core Ideas, One by One

The harm principle: where exactly is the line

This is the foundation of the book and the most quoted — and most misquoted — sentence in it. Mill's formulation, paraphrased: the sole end for which mankind are warranted in interfering with the liberty of action of any of their number is self-protection; the only purpose for which power can be rightfully exercised over any member of a civilised community, against his will, is to prevent harm to others. His own good, either physical or moral, is not a sufficient warrant.

To see how narrow that is, look at what it rules out. It rules out very nearly every reason people actually give when they interfere with each other: because it's better for you, because it's wiser, because it's right, because we can't stand watching. Each is a fine reason to argue, persuade, entreat. None is a reason to compel. The line between remonstrance and punishment is the whole of Mill's political design.

One clause is usually skipped and shouldn't be: Mill explicitly refuses the natural-rights route. He says he will claim no advantage from any abstract right independent of utility, and treats utility as the final appeal — but "utility in the largest sense, grounded on the permanent interests of man as a progressive being" (paraphrased). He is not saying liberty is mine by birth and you may not touch it. He is saying a society that lets people try things out does better over the long run. Resting the case for freedom on consequences is what makes the book so persuasive (you need not accept any metaphysics first) and also what makes it fragile (what if someone produces evidence that suppression works better?).

Mill grants himself two exceptions. The first is uncontroversial: children, and those still under legal care — the principle applies to human beings "in the maturity of their faculties." The second is the great stain on the book: despotism, he writes, is a legitimate mode of government in dealing with "backward states of society" whose people are still in their nonage, provided the end is their improvement (paraphrased). Having just argued that you cannot decide for an adult what is good for him, he classifies an entire subcontinent as minors. The bill for that comes due in the objections section below.

Tyranny of the majority: what he is actually afraid of

Mill takes the phrase from Tocqueville, who had argued in Democracy in America that the danger in a democracy is not a despot but the crushing weight of majority opinion. Mill then pushes it a long step further: his subject is no longer bad laws passed by majorities but "social tyranny" — pressure applied with no statute, no trial, no official at all, purely through opinion and custom. His verdict is sharp: this kind of tyranny is worse than most political oppression, because although it seldom has extreme penalties behind it, it leaves fewer means of escape, penetrates far more deeply into the details of life, and enslaves the soul itself (paraphrased).

Why should that be? Political oppression is external. You know where the knife is; you can fear it, dodge it, hate it, and the part of you that hates it is still yours. Social pressure is internalised: it doesn't have to arrest you, it gets you to strangle the thought yourself. You start pricing the reaction before you speak; after a while you are not biting your tongue, you are not having the thought — it never gets born. And most people, asked whether something ought to be permitted, don't feel any need for a reason: "I don't like it" is promoted automatically into "it shouldn't be allowed."

He cites a live English example. Campaigners for prohibition claimed a "social right" — your drinking degrades the moral and social atmosphere I have to live in, and therefore violates my rights. Mill's response is that the principle is monstrous precisely because it has no stopping point: once "your way of living makes me uncomfortable" counts as an injury to me, no liberty whatever can survive it — what you eat, what you believe, whom you marry, what you do on Sundays, all fall to the same logic. Nearly every present-day argument about whether something should be banned is a fresh edition of that claimed social right.

Why you must tolerate an opinion you are sure is wrong: the four-branch argument

Chapter 2 is the best-built and most durable part of the book. Mill doesn't sloganeer; he does an exhaustion proof. Whether the suppressed opinion is true or false, suppressing it costs you something. He splits the cases four ways and closes each one.

Mill's four branches: the bill for silencing an opinion falls due whether the opinion is right or wrong.

The first branch turns on the assumption of infallibility. Mill anticipates the obvious reply — we never claimed to be infallible, only certain. His answer is hard: certainty is not the same as being right; you may act on your convictions, but you may not close the question on behalf of everyone alive and everyone yet to be born. Then he goes looking for the examples least convenient to himself: Socrates and Jesus were both put to death, and not by bad men. The sharpest case is Marcus Aurelius — in Mill's own estimate about as good a man as the ancient world produced, author of the Meditations — who nevertheless authorised the persecution of Christians, because he sincerely believed Christianity would dissolve the bonds holding society together. If he could get it wrong, nobody gets to offer "I meant well and I was sure" as grounds for silencing anyone. The usual recipe for persecution is not malice plus stupidity; it is good intentions plus confidence.

The fourth branch stings hardest today. A doctrine tends to begin dying at the moment it wins: people still profess it, still nod on Sundays, but it no longer enters a single actual decision. Mill's example is Christianity — maxims like the camel and the eye of the needle are universally assented to and almost never allowed to touch anyone's behaviour. The creed survives; the force is gone. What keeps a belief alive is not how often it is affirmed but how often it is seriously attacked. Hence his admiration for the Catholic Church's advocatus diaboli, the devil's advocate: an office whose whole job, in a canonisation, is to assemble the case against.

Mill also drew the limit on speech more clearly than many who quote him realise. The opinion that corn dealers starve the poor should go unmolested in the press, he says, but may justly be punished when delivered to an excited mob assembled outside a corn dealer's house (paraphrased). The difference is not in the content of the opinion but in whether the utterance is itself an act directed at a specific person's ruin. Every subsequent law about incitement descends from that paragraph.

Individuality and experiments of living: freedom is not just being left alone

Chapter 3 is the part readers skim and the part Mill cared about most. The first two chapters say others should not interfere; this one says why it matters. His claim is much stronger than ordinary tolerance: individuality is not the price of a free society, it is one of the elements of well-being.

He gives an image nobody has improved on: human nature is not a machine to be built after a model and set to do exactly the work prescribed for it, but a tree, which requires to grow and develop itself on all sides according to the tendency of the inward forces that make it a living thing (paraphrased). A machine is judged from outside — did it execute the designer's intention accurately? A tree is judged from inside — did it grow into its own complete shape? Treat a person as a machine and the best available outcome is "satisfactory." Treat a person as a tree and there is such a thing as becoming who you are. The book's epigraph is from Wilhelm von Humboldt: the end of man is the highest and most harmonious development of his powers into a complete and consistent whole.

Then the definition of character: a person whose desires and impulses are his own — the expression of his own nature, developed and modified by his own culture — is said to have a character; one whose desires and impulses are not his own has no character, no more than a steam engine has a character (paraphrased). Notice what Mill is doing here: he is defending desire and impulse. The moral common sense of his day (and ours) says strong desires are dangerous and the good person is the one who wants little. Mill disagrees — strong impulses are simply more energy, and energetic people are capable of more good and more harm alike; what settles the question is whether those impulses are organised by conscience and whether they are his. Flattening a person's energy does not produce a saint. It produces someone docile and empty.

From this comes experiments of living, argued on two levels: for the individual, because only he has access to his own feelings and circumstances, so everyone else is guessing; and — more important — for everyone, because nobody knows in advance which way of living is better, so somebody has to actually run the trials. Diversity is not decoration; it is a society's only experimental plot. A society that permits one way of living has not become wicked, it has gone blind. Hence the deliberately abrasive line: precisely because the tyranny of opinion makes eccentricity a reproach, it is desirable that people should be eccentric, in order to break through it — and that so few now dare to be eccentric marks the chief danger of the time (paraphrased). His name for the whole condition is the despotism of custom: the standing hindrance to human advancement, requiring no one to give an order, running entirely on this is how it has always been done.

Honesty requires a note here: his negative example is China — a people of great talent who learned very early the art of making everyone as much alike as possible, and became stationary as a result (paraphrased). The judgement is unfounded historically and inconsistent with his own principle: frightening Englishmen with an imagined static East is exactly the substitution of a group image for actual persons that he elsewhere condemns. The frame of the argument still holds; this piece of evidence is bad.

Self-regarding and other-regarding: does the line survive?

Chapter 4 takes on the obvious objection: no man is an island — how can your self-destruction fail to injure others? Your drinking wrecks your family; your idleness is paid for by the rest of us; your dissipation sets an example. If that objection works, the harm principle has no edges at all.

Mill's reply is the most technical move in the book and the one most often missed: he does not deny that self-harm affects others, he requires that the effect consist in the violation of a distinct and assignable obligation. The weight is on assignable: not a diffuse claim that society suffers, but a demand that you name the party and name the duty. The same act changes character depending on where it sits:

Mill's line does not ask whether the conduct is admirable. It asks whether there is a victim you can name.

Three things are explicitly barred from counting as harm. First, offence and disgust — and here he produces a beautiful example: a Muslim's revulsion at the eating of pork is sincere, deep, close to physical, in no way pretended; yet even in a country where Muslims are an overwhelming majority, forbidding others to eat pork would still be tyranny. "Large numbers of people are genuinely disgusted" never equals "someone has been harmed." Second, moral disapproval — that you consider him degraded is a judgement of yours, not a debt of his. Third, the merely contingent injury of a "bad example" — if indirect effects count, society may regulate everything.

But Mill never says self-regarding conduct carries no consequences. He separates punishment from natural penalty: you may cease to like a man who is ruining himself, may avoid him, may warn others what he is — these are not punishments but other people exercising their own liberty. What he forbids is organised coercion: law, fines, concerted boycott. Not other people's right to form a view of you. Today's arguments about cancellation manage to lose that distinction from both sides at once. On Mill's own frame both sides are wrong: private avoidance is liberty, organised pressure is social tyranny — and the line between them is the hardest thing in the book to draw.

The exceptions Mill grants himself: poisons, drink taxes, and the slavery contract

Chapter 5 applies the principle to cases, and shows how restrained a thinker he is — every exception exposes an edge of the rule.

Poisons: not banned, but labelled, registered, documented. A ban would destroy countless legitimate uses (poisons are medicines and industrial inputs). But society may require the seller to record the buyer, because such records are what he calls pre-appointed evidence (paraphrased) — they obstruct nobody's choice, they merely make a later crime traceable. The template still works: regulate the traceability of risk, not the choice. The same logic produces a finer line elsewhere: taxing drink for revenue is legitimate; taxing it in order to stop people drinking is not — if the purpose is to price the choice out of reach, the tax is a prohibition in disguise, paternalism wearing a fiscal hat.

The unsafe bridge: you may stop the man, but only long enough to tell him. If someone is about to cross a bridge known to be unsafe and there is no time to warn him, an official or a passer-by may seize him and turn him back without any real infringement of his liberty — because liberty consists in doing what one desires, and he does not desire to fall into the river (paraphrased). This is the ancestor of soft paternalism: the warrant for intervening is not "I know better than you what's good for you" but "you lack a fact; here it is; now it's yours to decide." Everything in modern behavioural economics that goes by the name of nudging came in through that door.

The most interesting exception: a person may not sell himself into slavery, and such a contract is void. The reason is not the intuition that slavery is bad but a piece of pure logic: we let a person dispose of himself freely because that is how he pursues his own ends — and the man who sells himself has permanently abdicated all future pursuit of any end whatever. "The principle of freedom cannot require that he should be free not to be free. It is not freedom to be allowed to alienate his freedom" (paraphrased). That is a criterion for every voluntary irreversible surrender, and it remains the sharpest blade available in arguments about addiction, extreme contracts and technological dependence.

One more must be stated plainly, because it makes modern readers wince: Mill holds that bringing a child into the world without the means to support it is a moral crime, and expressly approves the laws in some continental countries forbidding marriage unless the parties can show they can maintain a family. The man who has just argued that the state may not decide how an adult should live here issues a compulsory verdict on the most private matter there is. It exposes something real: the moment the harm principle admits harm to the not-yet-born, its boundary goes soft.

The last sentence of the book reads better in context: the worth of a state, in the long run, is the worth of the individuals composing it; a state that dwarfs its men — even for beneficial purposes, even to make them more docile instruments — will find that with small men no great thing can really be accomplished (paraphrased).

Mill Beyond This Book

On Liberty leaves its own foundation deliberately blank: Mill says in the text that he appeals not to abstract right but to "utility in the largest sense" — and what that enlarged utility is, he explains elsewhere. Read this book alone and it is easy to mistake him for an anti-utilitarian.

Utilitarianism (1863) — the surgery he performed on Bentham's system. Bentham held that pleasures differ only in quantity; Mill insists they differ in quality, with a test: where people who have fully experienced both pleasures consistently prefer one, that one is higher. Hence the famous line: better to be Socrates dissatisfied than a fool satisfied (paraphrased). Fold that back into On Liberty and the logic closes: he defends individuality and experiment so fiercely because, in his own calculus, a person capable of choosing and unfolding is itself the highest kind of good. Liberty is not a means to happiness; liberty is a component of it.

The Subjection of Women (1869) — the harm principle applied where he thought the deepest unfreedom lay, and precisely where On Liberty barely goes: inside the family. Two moves. First, existing differences between the sexes prove nothing, because nobody has ever seen a woman who was not shaped by coercion — "what is now called the nature of women is an eminently artificial thing" (paraphrased). Second, a marriage law that subordinated a wife's person and property to her husband made the household the last and most intimate school of despotism, where every boy learns that domination by birthright is natural. Legal marriage, he wrote, was the only remaining form of lawful slavery in England (paraphrased). In 1867, as an MP, he moved to replace "man" with "person" in the Reform Bill — the first formal proposal of women's suffrage in Parliament. It was defeated.

Principles of Political Economy (1848) — the forgotten side: Mill was no doctrinaire of laissez-faire. He drew a consequential distinction — the laws of production are natural laws, but the rules of distribution are human institutions and can be changed (paraphrased) — and in later life grew steadily more sympathetic to cooperatives and to some form of socialism.

The Argument in Skeleton

Step one: change the question. Discussions of liberty used to ask how much power a ruler should have. Once the ruler is elected, "the people exercising power over themselves" turns out to be false — the power is always exercised by a majority and always lands on a minority. So the real question becomes: how much power should society have over an individual? The answer is a single word: harm. Not "for his own good," not "because it's right," and least of all "because we can't stand watching" — and the principle rests not on natural rights but on the claim that a society which permits trial and error does better over time.

Step two: hold the most important ground first — thought and discussion. Whatever the truth-value of a suppressed opinion, suppression costs something, and the sharpest cost is that an unchallenged truth rots into dogma. Then he converts the negative liberty into a positive reason: a person is a tree, not a machine, and diverse experiments of living are the only way the species ever discovers a better way to live. The chief enemy is therefore the despotism of custom, which needs no one to give an order.

Step three: close off "no man is an island," draw a workable line, then grant your own exceptions. Harm must consist in a violation of a distinct and assignable obligation; poisons may be sold but must be documented; drink may be taxed but not priced out of existence; the man on the unsafe bridge may be stopped long enough to be told; the slavery contract is void. Every exception points at the same criterion: intervention may target information and irreversibility, never "I think you chose wrongly." And it all lands on one sentence: the worth of a state is the worth of the people composing it, and a state that makes its people small will accomplish nothing large.

Misreadings and Serious Objections

Misreading one: Mill believed truth wins in open debate. He denied it flatly. The dictum that truth always triumphs over persecution, he wrote, is one of those pleasant falsehoods men repeat after one another until they pass into commonplaces, and which all experience refutes (paraphrased). His actual position is far colder: truth's only advantage is that it keeps being rediscovered until some era finally fails to suppress it. Incidentally, the "marketplace of ideas" is not Mill's phrase — it descends from Justice Holmes's 1919 dissent in Abrams v. United States. Attributing it to Mill swaps his careful argument for a cheerful slogan he would not have endorsed.

Misreading two: the harm principle means that as long as it's legal, nothing you say or do should have consequences. Mill himself separated punishment from natural penalty, and himself drew the line at incitement — what he protects you from is organised coercion, not other people's opinion of you. Misreading three: this is a book about the individual against the state. The introduction says the opposite outright: social tyranny is worse than political oppression because it penetrates deeper and offers fewer escapes. Read as an anti-government pamphlet, the book loses its more original half.

The strongest contemporary attack: James Fitzjames Stephen, Liberty, Equality, Fraternity (1873). Stephen, a criminal-law judge, argued that every society runs on coercion, that law and opinion have always enforced morality, and that the real question is only which coercions are worth their cost — not some single dividing principle. He also thought Mill naive about discussion: most people are moved by authority, habit and interest, not by argument. The force of the attack is that the harm principle looks like a boundary while quietly packing every hard question into the word "harm."

The great twentieth-century sequel: Devlin against Hart (1959–1963). The occasion was the Wolfenden Report of 1957, which recommended that consensual homosexual acts between adults cease to be crimes — a direct application of the harm principle. Lord Patrick Devlin objected on sociological grounds: a society is held together by a shared morality, the disintegration of that morality is the disintegration of the society, and so society may legislate against conduct that provokes the intolerance, indignation and disgust of the reasonable man. H. L. A. Hart replied in Law, Liberty and Morality along Millian lines: there is no evidence that moral disagreement dissolves societies, and treating majority disgust as a ground for law is exactly the elevation of taste into right that Mill warned about. Hart conceded one thing — some paternalism (seat-belt laws) is acceptable — while rejecting legal moralism, the criminalising of conduct merely because it is thought immoral. The practical outcome was not small: partial decriminalisation in England and Wales in 1967.

Structural objections: the self-regarding line is unstable, and the anti-paternalist premise may not hold. Almost any act has spillovers — pollution, contagion, insurance costs, the erosion of norms. And Mill's case against paternalism presupposes a well-informed adult with stable preferences, whereas present bias, addiction and engineered choice environments all make "his own choice" a doubtful phrase. Modern soft paternalism and nudging come in through the door Mill himself left open — the unsafe bridge. In fairness, none of this refutes the harm principle; it relocates the argument from "should you live that way" to "is there an assignable victim here" — much better ground, but it saves nobody the judgement call.

The colonial exception, which cannot be excused for him. "Despotism is a legitimate mode of government in dealing with barbarians, provided the end be their improvement" and his thirty-five years at East India House are two faces of one fact. The damage is logical as well as moral: once insufficient civilisation counts as a reason to withhold self-determination, the same reason is available against any population judged not yet grown up — and who does the judging has always been a question of power. The honest reading: the core argument remains formidable, and its author declined to apply it to everyone — which is the one failure this particular book could least afford.

Ten Sentences