This essay originally appeared in Christopher B. Gray (ed.), Philosophy of Law:  An Encyclopedia, Garland Pub. Co, 1999, II.632-635.   Copyright © 1999, Peter Suber.

Peter Suber, Philosophy Department, Earlham College

"Paternalism" comes from the Latin pater, meaning to act like a father, or to treat another person like a child. ("Parentalism" is a gender-neutral anagram of "paternalism".) In modern philosophy and jurisprudence, it is to act for the good of another person without that person's consent, as parents do for children. It is controversial because its end is benevolent, and its means coercive. Paternalists advance people's interests (such as life, health, or safety) at the expense of their liberty. In this, paternalists suppose that they can make wiser decisions than the people for whom they act. Sometimes this is based on presumptions about their own wisdom or the foolishness of other people, and can be dismissed as presumptuous. But sometimes it is not. It can be based on relatively good knowledge, as in the case of paternalism over young children or incompetent adults. Sometimes the role of paternalist is thrust upon the unwilling, as when we find ourselves the custodian and proxy for an unconscious or severely retarded relative. Paternalism is a temptation in every arena of life where people hold power over others: in childrearing, education, therapy, and medicine. But it is perhaps nowhere as divisive as in criminal law. Whenever the state acts to protect people from themselves, it seeks their good; but by doing so through criminal law, it does so coercively, often against their will.

Which acts should be criminalized and which acts are none of the state's business? How far does one have a right to harm oneself, to be different, or to be wrong? To what extent should people be free to do what they want if others are not harmed? What is harm? When is consent free and knowing? When do we think clearly and wisely enough, and when are we sufficiently free of duress and indoctrination, to be left to follow our own judgment, and when should we be restrained by others? Who should restrain whom, and when? These are the questions raised by paternalism.

Before we examine the issues more closely, consider the very wide range of paternalistic legislation. Acts which are often prohibited by the criminal law, but which have been alleged by serious writers to be victimless or harmless, at least for consenting adults, include the following: riding a motorcycle without a helmet, gambling, homosexual sodomy, prostitution, polygamy, making and selling pornography, selling and using marijuana, practicing certain professions without a license (law, medicine, education, massage, hair-styling), purchasing blood or organs, suicide, assisting suicide, swimming at a beach without a lifeguard, refusing to participate in a mandatory insurance or pension plan, mistreating a cadaver, loaning money at usurious interest rates, paying a worker less than the minimum wage, selling a prescription drug without a prescription, aggressive pan-handling, nudity at public beaches, truancy, flag burning, duelling, ticket scalping, blackmail, blasphemy, and dwarf-tossing.

Paternalism protects people from themselves, as if their safety were more important than their liberty. By contrast, the harm principle, famously articulated by Mill, holds that limiting liberty can only be justified to prevent harm to other people, not to prevent self-harm. More precisely, coercion can only be justified to prevent harm to unconsenting others, not to prevent harm to which the actors competently consent.

The usual legal prohibitions of murder, rape, arson, and theft are not paternalistic, since these acts harm unconsenting others; for the same reason, criminal legislation in these areas is consistent with the harm principle. Legal paternalism and the harm principle come into conflict over (1) competent self-harm and risk of self-harm, (2) harm to consenting others, and (3) harmless acts. The harm principle demands that we tolerate all three types of act, but paternalists often wish to regulate them. If a competently consenting person is not a victim, then these three types act are victimless. Under the harm principle, victimless crimes must be decriminalized, and virtually all paternalism over competent adults ended. The harm principle creates a "zone of privacy" for consensual or "self-regarding" acts, within which individuals may do what they wish and the state has no business interfering, even with the benevolent motive of a paternalist.

But the harm principle does not bar all paternalism. It permits paternalism over the incompetent, such as young children, the retarded, and perhaps those whose ability to make decisions is compromised by ignorance, deception, duress, or clouded faculties. In these cases, the consent to self-harm is not competent and need not be respected. As we will see below, the harm principle also permits what might be called self-paternalism or consensual paternalism.

Every legal system known to us seems to have some paternalistic criminal prohibitions. Conversely, the harm principle has apparently never been embraced without qualification by the laws of any country. If we wish to limit legal paternalism with a principle, the harm principle is the leading candidate. However, even informed proponents of the principle are far from agreement on (1) which acts harm only the actor, (2) which consents are valid, and (3) which acts are harmless. Finally, (4) if "harm" is defined broadly, or "valid consent" narrowly, then even the harm principle will fail to provide a meaningful zone of privacy or barrier to paternalism. Let us look more closely at these issues.

When does an act harm only the actor? Informed people disagree on whether the valid consent of recreational drug users, or truants, covers all the people likely to be harmed by drug use or truancy. If an act harms others, can we be sure that it only harms consenting others? This can be difficult to ascertain, especially if we concede with Mill that every act "affects" everyone, if only indirectly and remotely. A motorcycle rider who consents to the risks of riding without a helmet, and who suffers traumatic head injury, may harm many people who did not consent, e.g. his emotional and financial dependents, fellow members of his insurance pool, and taxpayers who support highway patrols, ambulance services, and public hospitals. If increasing my taxes or insurance premiums harms me for the purposes of the harm principle, then I might be harmed by the act which the motorcycle rider thought was private and self-regarding. This special application of the harm principle is called the "public charge argument" for coercion. It is not paternalistic, since it is directed against harm to unconsenting others, not against self-harm. If we can prohibit riding a motorcycle without a helmet because of the harmful "public charge" it levies on unconsenting others, then we can prohibit eating fatty foods on the same grounds. In a welfare state which shifts costs to compensate those who harm themselves, virtually all self-harm will be other-harm too; hence virtually every corner of life could be regulated by law without violating the harm principle, and virtually all paternalism would be justified.

When is consent valid? Duelling was outlawed in large part because lawmakers believed that even those who seemed to consent to a duel were giving invalid consents procured through extreme pressure and duress. Today one hears informed people disagree on whether prostitutes, drug addicts, indigent buyers of lottery tickets, workers willing to take less than the minimum wage, and students willing to have sex with their professors, are giving valid consents.

What is harm? Is public nudity harmful? Is the peddling of quack remedies for cancer harmful? Is divorce? Television violence? Well-funded commissions and independent social scientists disagree on whether pornography tends to harm women as a class. Liberals and radicals disagree on whether offended sensibilities are a kind of harm. Is harm by omission harm in the relevant sense? If I refuse to stop by a highway accident to render aid, or if I refuse to donate a kidney, have I caused harm? If these acts and omissions are harmless, then to prohibit them is paternalism (or legal moralism); if they are harmful, then to prohibit them is justified by the harm principle.

Sometimes a legislature will prohibit an act while conceding that the act can be harmless and the consent valid. For example, sodomy is still outlawed in many places even for consenting adults in private. Here the issue is not consent or harm, or the effect on the unconsenting public, but the morality of the act as such. To prohibit a harmless act solely on moral grounds is a special way acting for people's own good and making their consent irrelevant; this makes it a special form of paternalism. It is usually called "legal moralism".

Perhaps paternalism by legislators over young children and incompetent adults is as justified as paternalism over the same individuals by their parents. If so, then we must decide who is "young" and who is "incompetent" for the purposes of law. Should we use flat age cutoffs, as we do for driving automobiles and drinking alcohol? Or should we use one-on-one interviews with experts, as we do for competency to stand trial and involuntary civil commitment? Age cutoffs are administratively convenient, but they are based on presumptions which we know will be false in a foreseeable number of cases; and to apply them when false will be unjust. Careful interviews minimize these problems, but at such a great cost that many utilitarians find it prohibitive. Moreover, it is not at all clear that careful interviews can satisfactorily identify competency, since competency (in this context) is as much a political question as a medical one.

The harm principle holds that competent consents should take priority over benevolent legislative limits on our liberty. Paradoxically, this entails support for what might be called consensual paternalism or self-paternalism. If I make a living will when of sound mind, asking to be coerced for my own good in certain ways if I should ever become incompetent, then I am paternalizing myself, or consenting to a regimen in which others paternalize me. For this reason it is less objectionable than classical paternalism.

In a democracy, paternalism in the criminal law can to some extent be construed as self-paternalism. If "we" made the laws against usury and gambling, then "we" are restraining only ourselves. Before we justify these laws as self-paternalism, however, we must ask whether we are describing our democracy accurately or platitudinously. If laws to protect citizens from themselves were made by one non-representative faction, class, or bloc, or if the electoral process is distorted so that the outcomes of elections do not represent true social consent, then we may be dealing less with consensual self-paternalism than with majoritarian (or even minoritarian) tyranny. To overlook this possibility would justify paternalism by turning a blind eye to one of its most objectionable features.

If the legislature wishes to prohibit riding a motorcycle without a helmet, it may have a paternalistic or non-paternalistic rationale. If it believes the act is self-regarding, then it is being paternalistic; if it accepts the public charge argument, then it avoids paternalism and acts under the harm principle. There are many other ways to do what the paternalist does but without paternalism: notably, to widen the definition of harm, and to narrow that of valid consent. This fact, however, does not make arguments for and against paternalism vacuous. First, these arguments help articulate our general theory of justice, for example, by making clear that if an act harms only those who competently consent, then it must be tolerated. Second, we should not overestimate our freedom to rationalize here. Paternalism can be converted to non-paternalism only when we modulate the notions of harm and consent sufficiently. While this is sometimes distressingly easy, at least as often it is an exercise in sophistry, oversimplification, or self-deception.

Airaksinen, Timo. Ethics of Coercion and Authority: A Philosophical Study of Social Life. Pittsburgh: University of Pittsburgh Press, 1988.

Buchanan, Allen, and Dan Brock. Deciding for Others: The Ethics of Surrogate Decision-making. Cambridge: Cambridge University Press, 1990.

Devlin, Patrick. The Enforcement of Morals. Oxford: Oxford University Press, 1965.

Dworkin, Gerald. The Theory and Practice of Autonomy. Cambridge: Cambridge University Press, 1988.

Feinberg, Joel. The Moral Limits Of The Criminal Law. Oxford: Oxford University Press, 1984-1988. Vol. I: Harm to Others (1984); Vol. II: Offense to Others (1985); Vol. III: Harm to Self (1986); Vol. IV: Harmless Wrongdoing (1988).

Grey, Thomas C. The Legal Enforcement of Morality. New York: Alfred A. Knopf, 1983.

Hart, H.L.A. Law, Liberty and Morality. Stanford: Stanford University Press, 1963.

Hodson, John D. The Ethics of Legal Coercion. Dordrecht: D. Reidel, 1983.

Jordan, Shannon M. Decision Making for Incompetent Persons: The Law and Morality of Who Shall Decide. Springfield, Illinois: Charles C. Thomas, 1985.

Mill, John Stuart. On Liberty. Indianapolis: Hackett Pub. Co., 1978. (Original, 1859.)

Rosenbaum, Alan S. Coercion and Autonomy: Philosophical Foundations, Issues, and Practices. Westport, Conn.: Greenwood Press, 1986.

Sartorius, Rolf (ed.) Paternalism. Minneapolis: University of Minnesota Press, 1983.

VanDeVeer, Donald. Paternalistic Intervention: The Moral Bounds on Benevolence. Princeton: Princeton University Press, 1986.

[Blue Ribbon] Peter Suber, Department of Philosophy, Earlham College, Richmond, Indiana, 47374, U.S.A. Copyright © 1999, Peter Suber. Copyright © 1999, Peter Suber.