Libertarianism and Legal Paternalism

by John Hospers
Department of Philosophy,
University of Southern California


In his book Principles of Morals and Legislation, the eighteenth-century philosopher and legislator Jeremy Bentham divided all laws into three kinds:

(1) laws designed to protect you from harm caused by other people;
(2) laws designed to protect you from harm caused by yourself; and
(3) laws requiring you to help and assist others. Bentham held that only the first kind of laws were legitimate; and in general libertarians would agree with him.

The third class of laws, sometimes called “good Samaritan” laws, are greatly on the increase today, and their principal examples are not laws requiring you to assist persons in trouble (such as accident victims) although these are on the increase,’ but rather laws-both Congressional and bureaucratic-having to do with income redistribution, such as welfare and food stamps and programs for the disadvantaged. Bentham argued persuasively against these laws as well; but he also condemned laws of the second kind, and it is these I propose to discuss in this paper. Legislation designed to protect people from themselves is called “paternal legislation,” and the view that such laws are legitimate and ought to be passed is called “legal paternalism.”


Legal moralism is the view that the entire nation should be governed by one morality and/or religion, with dissent from the official view being punishable as a crime. Examples of legal moralism are the Catholic Church prior to the Reformation and Iran under the Ayatollah Khomeini.

Legal paternalism is the view that the law should, at least sometimes, require people to act (a) against their will (b) for their own good, in that way protecting them from the undesirable consequences of their own actions.

The term derives from the Latin “pater” (father): just as a kind father protects his children against harm and danger, pulling the child away from the speeding car or from the precipice down which he is about to fall, so the State should protect its citizens, not only against harm inflicted on them by other citizens, but also against harm which they might inflict on themselves.

Thus, according to legal paternalists, the State should prohibit drugs because otherwise people might take them, and even if the danger is only to their own health or life the State should protect such values for them if they are too foolish or incompetent to do so for themselves. Or again, the State should protect people from their own profligacy by forced savings, such as social security.

Libertarians, of course, are vigorously anti-paternalistic, believing as they do that people should absorb the consequences of their own actions, and that in any case the State has no right to legislate what people should do as long as their actions harm no one else. The concept of “harm” is admittedly vague:l some people would say, for example, that a teacher is harming their children more by teaching them anti-Christian doctrines than by injuring their physical bodies, and if such people had their way they would impose not only legal paternalism but a whole system of legal moralism. Most Christians today, however, aware of what would happen if each moral or religious sect tried to impose its views on everyone in this way, would resort to persuasion rather than to force, and however evil they might find certain teachings to be they would stop short of wanting them declared illegal. But disagreement about what constitutes harm continues: some consider X-rated movies harmful, others say the same about nude beaches, and still others would make the same assertion about certain theories of education. Yet most of those who say this (in the case of education, at least, often with good reason) would stop short of saying that those who inflict this alleged harm should be subject to civil or criminal prosecution. “Harm” is usually construed by libertarians, in accordance with their own political philosophy, to include

(a) bodily injury, such as assault and battery,
(b) damage to or theft of property, and
(c) violation of contract; and
accordingly it is only these that libertarians usually seek to prohibit by law.


Even libertarians are not, however, opposed as a rule to all paternalism. There are several groups of people in behalf of whom some degree of paternalistic action would be considered proper.

1. Infants and children. Infants cannot take care of themselves at all, and children cannot in many ways. Children do make decisions, hut lacking experience they often fail to comprehend the consequences of their own proposed actions. Views on children’s rights are a hotbed of current controversy; but there is probably no parent who has not at some time used coercion in order to prevent some harm to the child or bring about some good.A degree of paternalism concerning children is also embodied in the legal system: for example, if parents demonstrably abuse their children, the State takes the children out of the parent’s custody for the children’s own good, even if such action may not be in accord with the children’s own wishes at the time. The rationale of this is that the parents have proved themselves to be unfit custodians of the children’s rights.

2. The senile. When an elderly couple can no longer take care of themselves but refuse to leave their home, and when they consistently refuse to pay the utility bills and the heat and light are cut off, it is customary for a near relative to obtain power of attorney from the court in order to pay the bills and perhaps conduct other business transactions on behalf of the parents even if the parents are unwilling, in order to protect the parents from the consequences of their own actions. Though there has been little discussion of this, it is probable that most libertarians would go along with a degree of paternalism in such cases; at least it would bespeak a certain crassness to say, “If they’re so stupid or forgetful as not to pay their utility bills, let them freeze!” Our ordinary assumption is that people are able to estimate to some extent the probable consequences of their own actions, and this assumption is unjustified in the case of senility, just as it sometimes is in the case of children.

3. The mentally incompetent (a wider class than “the insane”). This is hardly a clear-cut group, but there are many people who are quite unable to functiqn in the world and quite as unable to fend for themselves as are young children. In most states people are at least temporarily institutionalized when they are “in imminent danger of harming themselves or others.” Libertarians in general are opposed to the compulsory institutionalization of persons who have committed no legal crimes; but it is not clear that all libertarians would be committed to opposing the non-voluntary incarceration of a knife-wielding psychotic in an aggressive phase when he was bent on killing the children in the neighborhood. Others might approve a person’s compulsory incarceration if he was a danger to himself, or even if he was simply unable to function, e.g., to know how to find food or shelter even if he had the money in his pocket.


But let us leave these groups aside for the moment. What about “ordinary normal adults”? At least, one would think, we should be totally opposed to any paternalism with respect to them. “Neither one person, nor any number of persons,” wrote John Stuart Mill in On Liberty, “is warranted in saying to another human creature of ripe years, that he shall not do with his life for his own benefit what he chooses to do with it. . . . The only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others. He cannot rightfully be compelled to do or forbear because it will be better for him to do so, because it will make him happier, because, in the opinion of others, to do so would be wise, or even right.”

Mill, a disciple of Bentham, was a utilitarian, and based his ethical conclusions on whatever was for “the greatest good of society”. But it is doubtful whether he could justify his strong anti-paternalism on utilitarian grounds. It may be that forcing motorcyclists to wear helmets for their own protection produces in its total consequences more good, e.g., more total happiness and less unhappiness, than the policy of not forcing them-partitularly if there are lots of careless riders.

It may even be that the policy of having parents arrange marriages produces less unhappiness than having young people (especially when they are emotionally immature) decide these matters for themselves; yet Mill would have them decide for themselves, even make their own mistakes and hopefully profit from them. In fact Mill, not in his Utilitarianism but in On Liberty, bases his anti-paternalistic stand on quite different considerations. “There is a part of the life of every person who has come to years of discretion, within which the individuality of that person ought to reign uncontrolled either by any other person or by the public collectively.” And again, from On Liberty, “A man’s mode of laying out his existence is the best, not because it is best in itself, but because it is his own mode. . . .It is the privilege and proper condition of a human being, arrived at the maturity of his faculties, to use and interpret experience in his own way.” Mill here is “saying something about what it means to be a person, an autonomous agent. It is because coercing a person for his own good denies this status as an independent entity that Mill objects to it so strongly and in such absolute terms. To be able to choose is a good that is independent of the wisdom of what is chosen.”3

The question I now want to ask is, Are libertarians committed to being one hundred percent anti-paternalistic, leaving aside the groups described in the previous section?

We are sometimes paternalistic with non-deranged adults, and believe ourselves to be quite justified in being so. A friend or spouse says to you, “Be sure to get me up at 7 o’clock; my job depends on it. Force me if you have to. No matter what I say at the time, get me up.” If you do so, contrary to the person’s wishes at the time, do you as a libertarian feel guilt and remorse? No, because even though forcing him to get up at that time is contrary to his wishes as of that moment, it is in accord with his long-term goals for himself. We are in a position in which we have to sacrifice either his short-term goal (staying asleep) or his long-term goal (keeping his job), and we consider it preferable to honor his long-term goal.

The attendant at a hospital force-feeds a patient who needs nourishment in order to live but refuses to take it. Should the libertarian say “If he doesn’t want food, it’s wrong to force him to take it”-thus letting him die? Surely not. What we will do (or at the very least, may permissibly do) is to go counter to his present desires, which may last a day or a week, in order to fulfill his long-term desire (which was constant prior to his present illness), which was to remain alive. When the patient has recovered he may thank us for force-feeding him: “It saved my life.” If this happened, would the libertarian still say that the force-feeding was wrong? Even if we have no independent evidence at the time that the patient’s attitude was pro-life, we may tentatively infer this from the fact that he has already lived this long, and are justified in having a presumption that he wishes to live. If he is grateful to us for saving his life, this alone justifies our previous action; and if he still wants to die after his recovery, he is still alive to make that choice, and there remain many ways in which he can undertake to bring about his own death if he so chooses. Some decisions, once made, are extremely far-reaching, or dangerous, or irreversible-sometimes all three at once, as in the present case. When this is so, we act paternalistically on the person’s behalf, so that he can live to freely choose another day.


It is one thing to be justified in doing X; it is another thing to require everyone to do X by law. Is there any justification at all for legal paternalism?

Mill himself thought there were occasions when legal paternalism was justified. He held, for example, that a contract by which a person agrees to sell himself into perpetual slavery should be null and void-as indeed it would be declared by virtually any court in the Western world. But why, if a person signs such a contract, should anyone interfere with it? “The reason for not interfering, unless for the sake of others, with a person’s voluntary acts,” wrote Mill in On Liberty, “is consideration for his liberty. . . . By selling himself for a slave, he abdicates his liberty; he foregoes any future use of it beyond that single act. He therefore defeats, in his own case, the very purpose which is the justification of allowing him to dispose of himself. …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.” The reason for not honoring such a contract is the need to preserve the liberty of the person to make future choices. Paternalism is justified at time t, in order to preserve a wider range of freedom for that individual at times t2,t,, tn,

Perhaps this example is extreme, or at any rate unique. Let us return then to our more mundane example, the law (which exists in all the states of the United States except four) requiring cyclists to wear helmets for their own protection. But “for their own protection” is not the only reason why such laws have been passed. It is also for the protection of others-thus falling under the heading of impure paternalism rather than pure paternalism. (A law is purely paternalistic if it is solely for the individual’s protection; it is impurely paternalistic when it is partly for that reason and partly for other reasons.) Without a helmet, a cyclist involved in an accident is liable to get a permanent head injury, and under present welfare and disability laws he would be a permanent ward of the state, perhaps living on for decades at taxpayer expense. The Supreme Court of Rhode Island a few years ago upheld the helmet requirement on the ground that it was “not persuaded that the legislature is powerless to prohibit individuals from pursuing a course of conduct which could conceivably result in their becoming public charges.”

Committing suicide is commonly a criminal offense. (You can be killed for doing it.) Even unsuccessful attempts are punishable. Yet if your life is your own, haven’t you the right to take it whenever you wish? What right has the State to command you not to take it? None, we say. Yet the State orders its policemen, when a person tries to kill himself by jumping in the river, to do their best to rescue the would-be suicide provided they can do so without “substantial risk” to their own lives. Is there any justification at all for this rule? I believe that such a rule could be defended, for the kind of reason already given: by forcibly preventing a person from taking his life at time tl, he thereby enables the person to make his own choice later, whereas the person’s death would put an end to all future choices. Perhaps the person was in a depressed state of mind which would pass, if he lived; perhaps he was confused, or drugged, or deranged -the  oliceman is in no position to know when he sees the man jump. It is better to assume that in the long run the man wants to live, than to assume that his continuing and steady disposition (time t*, t2,. . . t,) would be to die. If one assumes that his attempt is only a temporary aberration, and acts accordingly, the rewards may be great; whereas if it is not merely a temporary aberration, but an abiding disposition, then the man will still be alive to make a choice for death at a later time.

Paternalism in such a case represents a kind of wager made by the person acting paternalistically on another’s behalf: “I’ll wager that the long-run trend of your desires is contrary to your apparent wish at the present moment, so I will act to preserve your long-term wish even if it means denying your present, and hopefully temporary, one.” In some cases it may even be justifiable, as in the case of teen-age marriages, to have an enforced waiting period: when the consequences of the act would be far-reacing and possibly catastrophic, it may be better to make the person wait or hesitate even if he doesn’t wish to at the time, just as one makes the person get up even if he doesn’t want to at the time.

An impulsive suicide leap would have far-reaching and irreversible consequences, so isn’t one justified in erring, if at all, on the side of caution? If the weeks go by and the person is still deeply depressed and refuses advice or therapy, then he can, with Marcus Aurelius, weigh the pros and cons carefully and still decide,  “The room is smoky, so I leave it.”


Rather than adopt the simplistic conclusion that all paternalistic action is wrong, I shall adopt a more moderate conclusion: I want to say that the greater the degree to which a person’s action (or a proposed action, or a thought-of action) is voluntary, the less are other persons (or institutions, especially the law) justified in behaving paternalistically toward that person.

But the key word here is “voluntary”. The popular conception of voluntariness, which is shared by most libertarians, seems to me only to skim the surface of the concept. The popular conception, embedded also in most libertarian literature, is that voluntariness means non-coercion. As long as you’ve not been coerced, this argument suggests, your decision is voluntary. But in my view much more than this is required.

1. Freedom from coercion and pressure. It is true, of course, that when coercion occurs the decision is not voluntary. But even here there are degrees. The limiting case of coercion is one in which, for example, someone stronger than you are forces your fingers around the trigger of the gun; you resist but without success. In that case it isn’t your act at all, but the act of the person who forced you. Still, you were coerced. More typically coercion consists not of overt physical action but of the threat of it: “If you don’t hand over your wallet, I’ll shoot.” Unlike the first case, in threat cases there is a choice: you can surrender your life instead of (or probably in addition to) your wallet. But it isn’t much of a choice, and handing over the wallet isn’t the choice we would have made except for the coercion-we were made to do something we would not voluntarily have done.

Threats, too, are a matter of degree. Threat of loss of life is more serious than threat of injury; threat of injury is (usually) more serious than threat of loss of employment; and a threat by your mother-in-law to move if you don’t do what she asks is still Less of a threat-indeed it may be not a threat at all, but rather its opposite, an inducement.’ Many libertarians are willing to call it coercion only if there is physical harm or threat of physical harm, but in my opinion this is much too narrow. A threat of loss of a job may not be much of a threat if you can easily obtain another; but if no others are obtainable within a hundred miles, or if your special skill is not one for which there is any longer much demand, or if you would have to move your whole family to another state, the threat of loss of a job could be very serious. In any case it’s not a job you would voluntarily have left- you would not have quit it hut for the coercion (and it is coercion, threatening the means by which you live, differing only in degree from threat to life or limb).

Indeed, any kind ofpressure put on you interferes with the voluntariness of your decision. The warden says, “If you don’t cooperate with us by joining the group therapy sessions, we’ll put you in the hole for two weeks.” Surely this compromises the voluntariness of the prisoner’s decision. Someone puts pressure on you to make a decision hastily when you wouldn’t have made it without the pressure; while this may not be comparable to loss of life or limb, it may seriously compromise the voluntariness of your decision. It may be that laws against duelling are justified because if duelling were legally permitted many people would feel great pressure to preserve their “macho” image by never turning down a challenge, and thus they are (not exactly forced, but) pressured (perhaps with enormous sociological pressure) into entering a duel time after time even though they would prefer not to, and would refrain but for the pressure. It’s not an outright case of coercion, but there is a continuum between coercion and pressure and when the pressure is of the kind I have described, an individual will be relieved and gratified, and in the long run fulfilling his life-plan much more in accordance with his own wishes, if the practice is prohibited by law. (Remember the film The Duellists, in which this kind of pressure ruins the protagonist’s whole life. How different is that from killing him outright?) There is a certain paternalistic wisdom in the remark of that eminent philosopher Groucho Marx in one of his films, when he wakes from a faint and says, “Force some brandy down my throat!”

Any influence, whether pressure or outright coercion, which keeps the process of decision making from “filtering through your mind” and thus triggers the decision with partial or no cooperation from your untrammeled decision-making faculties, tends to inhibit the full voluntariness of the decision. But freedom from coercion and pressure is only one of the conditions requisite for voluntary action.

2. Informed and Educated Consent. The decision must be informed, based on the facts relevant to the case, and purged of false information. If the merchant sells you what he says is a real diamond when it’s actually glass, and you pay the price of a diamond, your decision to pay is not voluntary: “You wouldn’t have paid that much voluntarily,” we say, at least not for a piece of glass. It’s not that you were coerced, or even pressured; you were defrauded, that is, you were fed false information in making your decision.

Fraud is only one special case. You think you are drinking water, it was water you asked for and your host at the party brought a clear liquid that looked like water, only it contained poison. Even though no pressure was placed upon you, it is not reasonable to hold that you are voluntarily drinking poison. Drinking the poison is not in these circumstances a voluntary act; drinking water would have been, but that is not what you are doing. Or: you start to walk across a bridge, not knowing that further down the bridge has collapsed (you can’t see it through the fog). You know that if it has collapsed you will likely fall to your death, but you don’t know that it has collapsed. Since your aim is to cross the bridge and not to commit suicide, your action, based on misinformation, is not voluntary. If a man really thought that when he jumped out of the 20th floor window he would float through the air, would his jumping to his death still be voluntary?

When a patient consents to participate in a medical experiment-he’s not threatened, not pressured-but some of the possible serious consequences or unpleasant side-effects of the experimental drug have been concealed from him, one would not say that he consented voluntarily to take the drug.There must not only be uncoerced consent, there must be informed consent. Because his consent is not informed, it is not fully voluntary. How informed must it be to be “really informed”? The general formula is: he must be told all the relevant facts prior to making his decision. But this too turns out to be a matter of degree: one could go on forever citing medical facts which might turn out to be relevant; can one ever be quite sure one has reached an end of citing such facts? Even if the physician or researcher has cited all the facts he knows, there may still he others he doesn’t know which are highly relevant to the patient’s decision, even to his life or death. It would seem, then, that a patient can have “informed consent” but not ‘yully informed consent.” If full (complete) information is required for voluntariness, the patient’s consent must always be something less than fully voluntary. But once again, this is a matter of degree.

When prisoners, or patients in mental hospitals, are encouraged to offer themselves as experimental guinea pigs, it is highly probable that there is, lurking in the background if not in the foreground, some external pressure (punishment if you don’t, reward if you do). But in addition to this, it is seldom indeed that the patient is told even all the relevant information that the physician knows; what happens is more like “How would you guys like to join us in an interesting experiment, which won’t take much of your time” and so on. Thus the consent fails of voluntariness in both counts.

It would hardly be an overstatement to say that the consent of children to participate in such an experiment can never be wholly voluntary, and that “voluntary consent,” though it may be required in such a case, can never be given. Even if the child could reel off all the information an unusually loquacious physician has presented regarding the new medication, the child is not in a position to appreciate the force of that information. How many children can really understand the full force of a simple statement like “There’s a 50-50 chance that you’ll die”? Children can make all kinds of confident assertions, wagers, and challenges, not knowing fully what they really mean. When the twelve-year-old is offered some L.S.D., with the invitation “It’ll give you a wonderful high,” he may accept it eagerly, just as a baby might play with a stick of dynamite or a loaded gun. For this reason, contrary to what some libertarians apparently believe, all such invitations by others should be prohibited by law, for the child’s protection.

The child cannot give informed consent, much less “educated consent”-and those who would take advantage of the child’s incapacity should be met with the full force of the criminal law. To say of the child that “after all he gave his consent” would be ludicrous if its consequences were not so tragic.

3. Healthy Psychological State. I believe that there is a third condition that must be fulfilled as well. A person may not be under coercion or outside pressure, and he may be fully informed of the relevant facts of the case, and yet he may make his decision in what I can only describe as an unsatisfactory-or irrational, depending on what that term is taken to mean-psychological state. A person may be mentally deranged; but lacking this extreme, he may be in a daze, or drugged, or in an acute state of grief or depression, or just simply confused.

Ordinarily when a person is in such a state he can hardly be described as “fully informed,” and so his action would fail of voluntariness by the second criterion. But there may well be occasions when he is not pressured and all of the facts are clearly before him, and yet he is in no position to make a decision such as he would make if he were not in such a psychological state. A person in a state of  depression might be quite lucid as to the facts, yet a recital of ordinarily horrifying facts, such as his own imminent death or the extinction of the entire world, may well not move him to any kind of action or response.

I do not wish to say that any decision we might label as unwise shows that the person is in such an “abnormal” psychological state; people can certainly act voluntarily and yet foolishly. I only wish to suggest that when a person is in such a mental state as I have indicated, his decisions should not be described as fully voluntary. A psychotic in a highly manic phase may jump out of a second-story window, quite without coercion and in full possession of information as to the probable effects of his action. It is primarily because of the mental state of such a person, not because of pressure or lack of information, that we hesitate to describe his actions as fully voluntary.

In discussing human action, libertarians place very great emphasis upon voluntariness. But in my opinion most libertarians conceive it too thinly. “If he was forced, he hasn’t acted voluntarilyH-this much libertarians all assent to. But too often they fail to see that voluntariness is not as simple as that -that once it is clear that no coercion or pressure has been applied, the action may yet fail of  being voluntary. I have argued that the simplistic conception of voluntariness not only fails to do justice to the concept, but is often highly unfortunate in its effects. And I have argued that voluntariness, like so many other concepts, is not a yes-or-no concept but a matter of degree: not only does coercion-pressure itself encompass a broad spectrum of influences, from the application of force at one end to the exertion of subtle psychological pressure on the other, but that even when no external pressure has been applied, an act may be only incompletely voluntary because of its failure to meet the other two conditions.


Whenever I have offered remarks in defense of paternalism in the previous pages, paternalistic action was to be taken in order to help a person achieve his own goals. The man wants to get up at 7 a.m. to keep his job, and by going against his 6 a.m. command we are helping him achieve what he himself (though not at that moment) wants. If a person’s suicidal impulse is transitory, we help him achieve his long-term goals- which all, of course, presuppose life-by not letting him kill himself now. Even when laws prohibiting duelling were defended, it was on the assumption that a life freed of this curse is what the person who is constantly being challenged to other duels really wants for himself.

But there is also paternalism which thwarts the person’s long-term goals. Laws limiting the number of hours per week a person may work are often defended as protecting that person; but what if the person doesn’t want any such thing? What if the person wants to work extra long hours this year in order to have money to start a possibly lucrative business next year? ”

But,” one may say, “surely laws or actions that thwart the person’s own goals can’t be paternalistic at all, because part of the definition of paternalistic action is that it’s for the person’s own good.” Yes, but there’s the rub: what is for the person’s good may not be the same as what he wants (even in the long run).  Suppose that what would be for his good is to develop his talents so as to have a fulfilling life, but that all he wants is to be a bum. Or suppose he is a drug addict, and all he wants for himself even over a life-span is a state of drug-soaked euphoria (he doesn’t mind if his life is short, as long as it is, by his own standards, sweet). Even if we believe, and even if we believe truly, that such a life does not serve his good-we think of the wasted talents and of what he might have achieved and enjoyed if he had not (on our view) thrown away his life-we are nevertheless faced with the fact that what we want for him is not the same as what he wants for himself.

Any kind of paternalism which consists of our acting against his will to achieve our goals for him, rather than our acting against his (present) will to achieve his own goals (assuming, of course, that he is sufficiently mature to have them), is the kind of paternalism which I believe libertarians should condemn. Libertarians have condemned all paternalism without recognizing its two distinct forms, one of which may sometimes be acceptable and the other not.

Once it is clear that our goals for a person do not coincide with his goals for himself, and once we have used reason and possibly persuasion to convince him (never force), and he still sticks to his own, then as libertarians we must conclude, “It’s his life, and I don’t own it. I may sometimes use coercion against his will to promote his own ends, but I may never use coercion against his will to promote my ends. From my point of view, and perhaps even in some cosmic perspective, my ideals for him are better than his own.

But his have the unique distinguishing feature that they are his; and as such, I have no right to interfere forcibly with them.” Here, as libertarians, we can stand pat. It is, after all, just another application of Kant’s Second Moral Law -that we should always treat others as ends in themselves, never as means toward our own ends.

I. See, for example, James Ratcliffe, ed., The Good Somariron and rhe Low (New York:Anchor Doubleday Books, 1966).
2. See Joel Feinberg, Social Philoso~hy (Englewood Cliffs, N.J.: Prentice-Hall vawrback. Foundations of ~hilosophy series, 1973), chaps. 2 and 3
3. Gerald Dworkin. “Paternalism.” The Monist 56. no. I.
4. See John Hospers, “Some Problems ~oncerni& Punishment and the Use of Force,” Reason (November 1972 and January 1973).

More from John Hospers