Neither one person, nor any number of persons 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.  (Mill)

I do not want to go along with a volunteer basis.  I think a fellow should be compelled to become better and not let him use his discretion whether he wants to get smarter, more healthy or more honest. (General Hershey)

I take as my starting point the "one very simply principle" proclaimed by Mill On Liberty...

That principle is, that the sole end for which mankind are warranted, individually or collectively, in interfering with the liberty of action of any of their number, is self-protection.  That the only purpose for which power can be rightfully exercised over any member of a civilized community, again 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.

This principle is neither "one" nor "very simple."  It is at least two principles; one asserting that self-protection or the prevention of harm to others is sometimes a sufficient warrant and the other claiming that the individual's own good is never a sufficient warrant for the exercise of compulsion either by the society as a whole or by its individual members.  I assume that no one, with the possible exception of extreme pacifists or anarchists, questions the correctness of the first half of the principle.  This essay is an examination of the negative claim embodied in Mill's principle--the objection to paternalistic interferences with a man's liberty.

I

By paternalism I shall understand roughly the interference with a person's liberty of action justified by reasons referring exclusively to the welfare, good, happiness, needs, interests or values of the person being coerced.  One is always well-advised to illustrate one's definitions by examples but it is not easy to find "pure" examples of paternalistic interferences.  For almost any piece of legislation is justified by several different kinds of reasons and even if historically a piece of legislation can be shown to have been introduced for purely paternalistic motives, it may be that advocates of the legislation with an antipaternalistic outlook can find sufficient reasons justifying the legislation without appealing to the reasons which were originally adduced to support it.  Thus, for example, it may be that the original legislation requiring motorcyclists to wear safety helmets was introduced for purely paternalistic reasons.  But the Rhode Island Supreme Court recently upheld such legislation on the grounds 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," thus clearly introducing reasons of a quite different kind.  Now I regard this decision as being based on reasoning of a very dubious nature but it illustrates the kind of problem one has in finding examples.  The following is a list of the kinds of interferences I have in mind as being paternalistic.

II

1. Laws requiring motorcyclists to wear safety helmets when operating their machines.
2. Laws forbidding persons from swimming at a public beach when lifeguards are not on duty.
3. Laws making suicide a criminal offense.
4. Laws making it illegal for women and children to work at certain types of jobs.
5. Laws regulating certain kinds of sexual conduct, for example, homosexuality among consenting adults in private.
6. Laws regulating the use of certain drugs when may have harmful consequences to the user but do not lead to antisocial conduct.
7. Laws requiring a license to engage in certain professions with those not receiving a license subject to fine or jail sentence if they do engage in the practice.
8. Laws compelling people to spend a specified fraction of their income on the purchase of retirement annuities (Social Security).
9. Laws forbidding various forms of gambling (often justified on the grounds that the poor are more likely to throw away their money on such activities than the rich who can afford to).
10. Laws regulating the maximum rates of interest for loans.
11. Laws against duelling

In addition to laws which attach criminal or civil penalties to certain kinds of action there are laws, rules, regulations, decrees which make it either difficult or impossible for people to carry out their plans and which are also justified on paternalistic grounds.  Examples of this are:

1. Laws regulating the types of contracts which will be upheld as valid by the courts, for example (an example of Mill's to which I shall return) no man may make a valid contract for perpetual involuntary servitude.
2. Not allowing assumption of risk as a defense to an action based on the violation of a safety statute.
3. Not allowing as a defense to a charge of murder or assault the consent of the victim.
4. Requiring members of certain religious sects to have compulsory blood transfusions.  This is made possible by not allowing the patient to have recourse to civil suits for assault and battery and by means of injunctions.
5. Civil commitment procedures when these are specifically justified on the basis of preventing the person being committed from harming himself.  The D. C. Hospitalization of the Mentally Ill Act provides for involuntary hospitalization of a person who "is mentally ill, and because of that illness, is likely to injure himself or others if allowed to remain at liberty." The term injure in this context applies to unintentional as well as intentional injuries.

All of my examples are of existing restrictions on the liberty of individuals.  Obviously one can think of interferences which have not yet been imposed.  Thus one might ban the sale of cigarettes, or require that people wear safety belts in automobiles (as opposed to merely having them installed), enforcing this by not allowing a motorist to sue for injuries even when caused by other drivers if the motorist was not wearing a seat belt at the time of the accident.

I shall not be concerned with activities which though defended on paternalistic grounds are not interferences with the liberty of persons, for example, the giving of subsidies in kind rather than in cash on the grounds that the recipients would not spend the money on the goods which they really need, or not including a $1,000 deductible provision in a basic protection automobile insurance plan on the ground that the people who would elect it could least afford it. Nor shall I be concerned with measures such as "truth-in-advertising" acts and Pure Food and Drug legislation which are often attacked as paternalistic but which should not be considered so.  In these cases all that is provided--it is true by the use of compulsion--is information which it is presumed that rational persons are interested in having in order to make wise decisions.  There is no interference with the liberty of the consumer unless one wants to stretch a point beyond good sense and say that his liberty to apply for a loan without knowing the true rate of interest is diminished.  It is true that sometimes there is sentiment for going further than providing information, for example when laws against usurious interest are passed preventing those who might wish to contract loans at high rates of interest from doing so, and these measures may correctly be considered paternalistic.

III

Bearing these example in mind, let me return to a characterization of paternalism.  I said earlier that I meant by the term, roughly, interference with a person's liberty for his own good.  But, as some of the examples show, the class of persons whose good is involved is not always identical with the class of persons whose freedom is restricted.  Thus, in the case of professional licensing it is the practitioner who is directly interfered with but it is the would-be patient whose interests are presumably being served.  Not allowing the consent of the victim to be a defense to certain types of crime primarily affects the would-be aggressor but it is the interests of the willing victim that we are trying to protect.  Sometimes a person may fall into both classes as would be the case if we banned the manufacture and sale of cigarettes and a given manufacturer happened to be a smoker as well.

Thus we may first divide paternalistic interferences into "pure" and "impure" cases. In "pure" paternalism the class of person whose freedom is restricted is identical with the class of persons whose benefit is intended to be promoted by such restrictions.  Examples: the making of suicide a crime, requiring passengers in automobiles to wear seat belts, requiring a Christian Scientist to receive a blood transfusion.  In the case of "impure" paternalism in trying to protect the welfare of a class of persons we find that the only way to do so will involve restricting the freedom of other persons besides those who are benefited.  Now it might be thought that there are no cases of "impure" paternalism since any such case could always be justified on nonpaternalistic grounds, that is in terms of preventing harm to others.  Thus we might ban cigarette manufacturers from continuing to manufacture their product on the grounds that we are preventing them from causing illness to others in the same way that we prevent other manufacturers from releasing pollutants into the atmosphere, thereby causing danger to the members of the community.  The difference is, however, that in the former but not the latter case the harm is of such a nature that it could be avoided by those individuals affected if they so chose.  The incurring of the harm requires, so to speak, the active cooperation of the victim.  It would be mistaken theoretically and hypocritical in practice to assert that our interference in such cases is just like our interference in standard cases of protecting others from harm.  At the very least someone interfered with in this way can reply that no one is complaining about his activities.  It may be that impure paternalism requires arguments or reasons of a stronger kind in order to be justified, since there are persons who are losing a portion of their liberty and they do not even have the solace of having it be done "in their own interest."  Of course in some sense, if paternalistic justifications are ever correct, then we are protecting others, we are preventing some from injuring others, but it is important to see the differences between this and the standard case.

Paternalism then will always involve limitations on the liberty of some individuals in their own interests but it may also extend to interferences with the liberty of parties whose interests are not in question.

IV

Finally, by way of some more preliminary analysis, I want to distinguish paternalistic interference with liberty from a related type with which it is often confused.  Consider, for example, legislation which forbids employees to work more than, say, forty hours per week.  It is sometimes argued that such legislation is paternalistic for if employees desired such a restriction on their hours of work they could agree among themselves to impose it voluntarily.  But because they do not the society imposes its own conception of their best interests upon them by the use of coercion.  Hence this is paternalism.

Now it may be that some legislation of this nature is, in fact, paternalistically motivated.  I am not denying that.  All I want to point out is that there is another possible way of justifying such measures which is not paternalistic in nature.  It is not paternalistic because, as Mill puts it in a similar context, such measures are "required not to overrule the judgment of individuals respecting their own interest, but to give effect to that judgment: they being unable to give effect to it except by concert, which concert again cannot be effectual unless it receives validity and sanction from the law" (Principles of Political Economy).

The line of reasoning here is a familiar one first found in Hobbes and developed with great sophistication by contemporary economists in the last decade or so.  There are restrictions which are in the interests of a class of persons taken collectively but are such that the immediate interest of each individual is furthered by his violating the rule when others adhere to it.  In such cases the individuals involved may need the use of compulsion to give effect to their collective judgment of their own interest by guaranteeing each individual compliance by the others.  In these cases compulsion is not used to achieve some benefit which is not recognized to be a benefit by those concerned, but rather because it is the only feasible means of achieving some benefit which is recognized as such by all concerned.  This way of viewing matters provides us with another characterization of paternalism in general.  Paternalism might be thought of as the use of coercion to achieve a good which is not recognized as such by those persons for whom the good is intended.  Again while this formulation captures the heart of the matter--it is surely what Mill is objecting to in On Liberty--the matter is not always quite like that.  For example, when we force motorcyclists to wear helmets we are trying to promote a good--the protection of the person from injury--which is surely recognized by most of the individuals concerned.  It is not that a cyclist doesn't value his bodily integrity; rather, as a supporter of such legislation would put it, he either places, perhaps irrationally, another value or good (freedom from wearing a helmet) above that of physical well-being or, perhaps, while recognizing that danger in the abstract, he either does not fully appreciate it or he underestimates the likelihood of its occurring.  But now we are approaching the question of possible justifications of paternalistic measures and the rest of this essay will be devoted to that question.

V

I shall begin for dialectical purposes by discussing Mill's objections to paternalism and then go on to discuss more positive proposals.

An initial feature that strikes one is the absolute nature of Mill's prohibitions against paternalism.  It is so unlike the carefully qualified admonitions of Mill and his fellow utilitarians on other moral issues. He speaks of self-protection as the sole end warranting coercion, of the individual's own goals as never being a sufficient warrant.  Contrast this with his discussion of the prohibition against lying in Utilitarianism:

Yet that even this rule, sacred as it is, admits of possible exception, is acknowledged by all moralists, the chief of which is where the withholding of some fact...would save an individual...from great an unmerited evil.

The same tentativeness is present when he deals with justice:

It is confessedly unjust to break faith with any one: to violate an engagement, either express or implied, or disappoint expectations raised by our own conduct, at least if we have raised these expectations knowingly and voluntarily.  Like all the other obligations of justice already spoken of, this one is not regarded as absolute, but as capable of being overruled by a stronger obligation of justice on the other side.

This anomaly calls for some explanation.  The structure of Mill's argument is as follows:

1. Since restraint is an evil the burden of proof is on those who propose such restraint.
2. Since the conduct which is being considered is purely self-regarding, the normal appeal to the protection of the interests of others is not available.
3. Therefore we have to consider whether reasons involving reference to the individual's own good, happiness, welfare, or interests are sufficient to overcome the burden of justification.
4. We either cannot advance the interests of the individual by compulsion, or the attempt to do so involves evils which outweigh the good done.
5. Hence the promotion of the individual's own interest does not provide a sufficient warrant for use of compulsion.

Clearly the operative premise here is (4), and it is bolstered by claims about the status of the individual as judge and appraiser of his welfare, interests, needs, et cetera:

With respect to his own feelings and circumstances, the most ordinary man or woman has means of knowledge immeasurable surpassing those that can be possessed by any one else.

He is the man most interested in his own well-being: the interest which any other person, except in cases of strong personal attachment, can have in it is trifling, compared to that which he himself has.

These claims are used to support the following generalizations concerning the utility of compulsion for paternalistic purposes.

The interferences of society to overrule his judgment and purposes in what only regards himself must be grounded on general presumptions; which may be altogether wrong, and even if right, are as likely as not to be misapplied to individual cases.

But the strongest of all the arguments against the interference of the public with purely personal conduct is that when it does interfere, the odds are that it interferes wrongly and in the wrong place.

All errors which the individual is likely to commit against advice and warning are far outweighed by the evil of allowing others to constrain him to what they deem his good.

Performing the utilitarian calculations by balancing the advantages and disadvantages, we find that: "Mankind are greater gainers by suffering each other to live as seems good to themselves, than by compelling each other to live as seems good to the rest." Ergo (4).

This classical case of a utilitarian argument with all the premises spelled out is not the only line of reasoning present in Mill's discussion.  There are asides, and more than asides, which look quite different and I shall deal with them later.  But this is clearly the main channel of Mill's thought and it is one which has been subjected to vigorous attack from the moment it appeared--most often by fellow utilitarians.  The link that they have usually seized on is as Fitzjames Stephen put is in Liberty, Equality, Fraternity, the absence of proof that the "mass of adults are so well acquainted with their own interests and so much disposed to pursue them that no compulsion or restraint put upon them by any others for the purpose of promoting their interest can really promote them."  Even so sympathetic a critic as H. L. A. Hart is forced to the conclusion that:

In Chapter 5 of his essay [On Liberty] Mill carried his protests against paternalism to lengths that may now appear to us as fantastic. ...No doubt if we no longer sympathise with his criticism this is due, in part, to a general decline in the belief that individuals know their own interest best.

Mill endows the average individual with "too much of the psychology of a middle-aged man whose desires are relatively fixed, not liable to be artificially stimulated by external influences; who know what he wants and what gives him satisfaction or happiness; and who pursues these things when he can."

Now it is interesting to note that Mill himself was aware of some of the limitations on the doctrine that the individual is the best judge of his own interests.  In his discussion of government intervention in general (even where the intervention does not interfere with liberty but provides alternative institutions to those of the market) after making claims which are parallel to those just discussed, for example, "People understand their own business and their own interests better, and care for them more, than the government does, or can be expected to do," he goes on to an intelligent discussion of the "very large and conspicuous exceptions" to the maxim that:

Most persons take a juster and more intelligent view of their own interest, and of the means of promoting it than can either be prescribed to them by a general enactment of the legislature, or pointed out in the particular case by a public functionary.

Thus there are things

of which the utility does not consist in ministering to inclinations, nor in serving the daily uses of life, and the want of which is least felt where the need is greatest.  This is peculiarly true of those things which are chiefly useful as tending to raise the character of human beings.  The uncultivated cannot be competent judges of cultivation.  Those who most need to be made wiser and better, usually desire it least, and, if they desire it, would be incapable of finding the way to it by their own lights.

...A second exception to the doctrine that individuals are the best judges of their own interest, is when an individual attempts to decide irrevocably now what will be best for his interest at some future and distant time.  The presumption in favor of individual judgment is only legitimate, where the judgment is grounded on actual, and especially on present, personal experience; not where it is formed antecedently to experience, and not suffered to be reversed even after experience has condemned it.

The upshot of these exceptions is that Mill does not declare that there should never be government interference with the economy but rather that

...in every instance, the burden of making out a strong case should be thrown not on those who resist but those who recommend government interference.  Letting alone, in short, should be the general practice: every departure from it, unless required by some great good, is a certain evil.

In short, we get a presumption, not an absolute prohibition.  The question is why doesn't the argument against paternalism go the same way?

I suggest that the answer lies in seeing that in addition to a purely utilitarian argument Mill uses another as well.  As a utilitarian, Mill has to show, in Fitzjames Stephen's words, that: "Self-protection apart, no good object can be attained by any compulsion which is not in itself a greater evil than the absence of the object which the compulsion obtains." To show this is impossible, one reason being that it isn't true.  Preventing a man from selling himself into slavery (a paternalistic measure which Mill himself accepts as legitimate), or from taking heroin, or from driving a car without wearing seat belts may constitute a lesser evil than allowing him to do any of these things.  A consistent utilitarian can only argue against paternalism on the grounds that it (as a matter of fact) does not maximize the good.  It is always a contingent question that may be returned by the evidence.  But there is also a noncontingent argument which runs through On Liberty. When Mill states that "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," he 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.  A man's "mode of laying out his existence is the best, not because it is the 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.

As further evidence of this line of reasoning in Mill, consider the one exception to his prohibition against paternalism.

In this and most civilised countries, for example, and engagement by which a person should sell himself, or allow himself to be sold, as a slave, would be null and void; neither enforced by law nor by opinion.  The ground for thus limiting his power of voluntarily disposing of his own lot in life, is apparent, and is very clearly seen in this extreme case.  The reason for not interfering unless for the sake of others, with a person's voluntary acts, is consideration for his liberty.  His voluntary choice is evidence that what he so chooses is desirable, or at least endurable, to him, and his good is on the whole best provided for by allowing him to take how own means of pursuing it.  But 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.  He is no longer free; but is thenceforth in a position which has no longer the presumption in its favour, that would be afforded by his voluntarily remaining in it.  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.

Now leaving aside the fudging on the meaning of freedom in the last line, it is clear that part of this argument is incorrect.  While it is true that future choices of the slave are not reasons for thinking that what he chooses then is desirable for him, what is at issue is limiting his immediate choice; and since this choice is made freely, the individual may be correct in thinking that his interests are best provided for by entering such a contract.  But the main consideration for not allowing such a contract is the need to preserve the liberty of the person to make future choices.  This gives us a principle--a very narrow one--by which to justify some paternalistic interferences.  Paternalism is justified only to preserve a wider range of freedom for the individual in question. How far this principle could be extended, whether it can justify all the cases in which we are inclined upon reflection to think paternalistic measures justified, remains to be discussed.  What I have tried to show so far is that there are two strains of argument in Mill--one straight-forward utilitarian mode of reasoning and one which relies not on the goods which free choice leads to but on the absolute value of the choice itself.  The first cannot establish any absolute prohibition but at most a presumption and indeed a fairly weak one given some fairly plausible assumptions about human psychology; the second, while a stronger line of argument, seems to me to allow on its own grounds a wider range of paternalism than might be suspected.  I turn now to a consideration of these matters.

VI  

We might being looking for principles governing the acceptable use of paternalistic power in cases where it is generally agreed that it is legitimate.  Even Mill intends his principles to be applicable only to mature individuals, not those in what he calls "non-age."  What is it that justifies us in interfering with children?  The fact that they lack some of the emotional and cognitive capacities required in order to make fully rational decisions.  It is an empirical question to just what extent children have an adequate conception of their own present and future interests but there is not much doubt that there are many deficiencies.  For example, it is very difficult for a child to defer gratification for any considerable period of time.  Given these deficiencies and given the very real and permanent dangers that may befall the child, it becomes not only permissible but even a duty of the parent to restrict the child's freedom in various ways.  There is however an important moral limitation on the exercise of such parental power which is provided by the notion of the child eventually coming to see the correctness of his parent's interventions.  Parental paternalism may be thought of as a wager by the parent on the child's subsequent recognition of the wisdom of the restrictions.  There is an emphasis on what could be called future-oriented consent--on what the child will come to welcome, rather than on what he does welcome.

The essence of this idea has been incorporated by idealist philosophers into various types of "real-will" theory as applied to fully adult persons.  Extensions of paternalism are argued for by claiming that in various respects, chronologically mature individuals share the same deficiencies in knowledge, capacity to think rationally, and the ability to carry out decisions that children possess.  Hence in interfering with such people we are in effect doing what they would do if they were fully rational.  Hence we are not really opposing their will, hence we are not really interfering with their freedom.  The dangers of this move have been sufficiently exposed by Berlin in his Two Concepts of Liberty.  I see no gain in theoretical clarity nor in practical advantage in trying to pass over the real nature of the interferences with liberty that we impose on others.  Still the basic notion of consent is important and seems to me the only acceptable way of trying to delimit an area of justified paternalism.

Let me start by considering a case where the consent is not hypothetical in nature.  Under certain conditions it is rational for an individual to agree that others should force him to act in ways which, at the time of action, the individual may not see as desirable.  If, for example, a man knows that he is subject to breaking his resolves when temptation is present, he may ask a friend to refuse to entertain his requests at some later stage.

A classical example is given in the Odyssey when Odysseus commands his men to tie him to the mast and refuse all future orders to be set free, because he knows the power of the Sirens to enchant men with their songs.  Here we are on relatively sound ground in later refusing Odysseus' request to be set free.  He may even claim to have changed his mind but, since it is just such changes that he wished to guard against, we are entitled to ignore them.

A process analogous to this may take place on a social rather than individual basis.  An electorate may mandate its representatives to pass legislation which when it comes time to "pay the price" may be unpalatable.  I may believe that a tax increase is necessary to halt inflation though I may resent the lower pay check each month.  However in both this case and that of Odysseus, the measure to be enforced is specifically requested by the party involved and at some point in time there is genuine consent and agreement on the part of those persons whose liberty is infringed.  Such is not the case for the paternalistic measures we have been speaking about.  What must be involved here is not consent to specific measures but rather consent to a system of government, run by elected representatives, with an understanding that they may act to safeguard our interests in certain limited ways.

I suggest that since we are all aware of our irrational propensities, deficiencies in cognitive and emotional capacities, and avoidable and unavoidable ignorance, it is rational and prudent for us to in effect take out "social insurance policies."  We may argue for and against proposed paternalistic measures in terms of what fully rational individuals would accept as forms of protection.  Now clearly, since the initial agreement is not about specific measures we are dealing with a more-or-less blank check and therefore there have to be carefully defined limits.  What I am looking for are certain kinds of conditions which make it plausible to suppose that rational men could reach agreement to limit their liberty even when other men's interest are not affected.

Of course as in any kind of agreement schema there are great difficulties in deciding what rational individuals would or would not accept.  Particularly in sensitive areas of personal liberty, there is always a danger of the dispute over agreement and rationality being a disguised version of evaluative and normative disagreement.

Let me suggest types of situations in which it seems plausible to suppose that fully rational individuals would agree to having paternalistic restrictions imposed upon them.  It is reasonable to suppose that there are "goods" such as health which any person would want to have in order to pursue his own good--no matter how that good is conceived.  This is an argument used in connection with compulsory education for children but it seems to me that it can be extended to other goods which have this character.  Then one could agree that the attainment of such goods should be promoted even when not recognized to be such, at the moment, by the individuals concerned.

An immediate difficulty from the fact that men are always faced with competing goods and that there may be reasons why even a value such as health--or indeed life--may be overridden by competing values.  Thus the problem with the Christian Scientist and blood transfusions.  It may be more important for him to reject "impure substances" than to go on living.  The difficult problem that must be faced is whether one can give sense to the notion of a person irrationally attaching weights to competing values.

Consider a person who know the statistical data on the probability of being injured when not wearing seat belts in an automobile and knows the types of gravity of the various injuries. He also insists that the inconvenience attached to fastening the belt every time he gets in and out of the car outweighs for him the possible risks to himself.  I am inclined in this case to think that such a weighing is irrational.  Given his life plans, which we are assuming are those of the average person, his interests and commitments already undertaken, I think it is safe to predict that we can find inconsistencies in his calculations at some point.  I am assuming that this is not a man who for some conscious our unconscious reasons is trying to injure himself nor is he a man who just likes to "live dangerously." I am assuming that he is like us in all the relevant respects but just puts an enormously high negative value on inconvenience--one which does not seem comprehensible or reasonable.

It is always possible, or course, to assimilate this person to creatures like myself.  I , also, neglect to fasten my seat belt and I concede such behavior is not rational but not because I weight the inconvenience differently from those who fasten the belts.  It is just that having made (roughly) the same calculation as everybody else, I ignore it in my actions.  [Note: a much better case of weakness of the will than those usually given in ethics tests.]  A plausible explanation for this deplorable habit is that although I know in some intellectual sense what the probabilities and risks are I do not fully appreciate them in am emotionally genuine manner.

We have two distinct types of situation in which a man act in a nonrational fashion.  In one case he attaches incorrect weights to some of his values; in the other he neglects to act in accordance with his actual preferences and desires.  Clearly there is a stronger and more persuasive argument for paternalism in the latter situation.  Here we are really not--by assumption--imposing a good on another person. But why may we not extend out interference to what we might call evaluative delusions?  After all, in the case of cognitive delusions we are prepared, often, to act against the expressed will of the person involved.  If a man believes that when he jumps out of the window he will float upwards--Robert Nozick's example--would not we detain him, forcible if necessary?  The reply will be that this man doesn't wish to be injured and if we could convince him that he is mistaken as to the consequences of his action, he would not wish to perform the action.  But part of what is involved in claiming that the man who doesn't fasten his seatbelts is attaching an incorrect weight to the inconvenience of fastening them is that if he were to be involved in an accident and severely injured he would look back and admit that the inconvenience wasn't as bad as all that.  So there is a sense in which, if I could convince him of the consequences of his action, he also would not wish to continue his present course of action.  Now the notion of consequences being used here is covering a lot of ground.  In one case it's being used to indicate what will or can happen as a result of a course of action and in the other it's making a prediction about the future evaluation of the consequences--in the first sense--of a course of action.  And whatever the difference between facts and values--whether it be hard and fast or soft and slow--we are genuinely more reluctant to consent to interferences where evaluative differences are the issue.  Let me now consider another factor which comes into play in some of these situations which may make an important difference in our willingness to consent to paternalistic restrictions.

Some of the decisions we make are of such a character that they produce changes which are in one or another way irreversible.  Situations are created in which it is difficult or impossible to return to anything like the initial state at which the decision was made.  In particular, some of these changes will make it impossible to continue to make reasoned choices in the future.  I am thinking specifically of decisions which involve taking drugs that are physically or psychologically addictive and those which are destructive of one's mental and physical capacities.

I suggest we think of the imposition of paternalistic interferences in situations of this kind as being a kind of insurance policy which we take out against making decisions which are far-reaching, potentially dangerous and irreversible,  Each of these factors is important.  Clearly there are many decisions we make that are relatively irreversible.  In deciding to learn to play chess, I could predict in view of my general interest in games that some portion of my free time was going to be preempted and that it would not be easy to give up the game once I acquired a certain competence.  But my whole life style was not going to be jeopardized in an extreme manner.  Further it might be argued that even with addictive drugs such as heroin one's normal life plans would not be seriously interfered with if an inexpensive and adequate supply were readily available.  So this type of argument might have a much narrower scope than appears to be the case at first.

A second class of cases concerns decisions which are made under extreme psychological and sociological pressures.  I am not thinking here of the making of the decision as being something one is pressured into--for example, a good reason for making duelling illegal is that unless this is done many people might have to manifest their courage and integrity in ways in which they would rather not do so--but rather of decisions, such as that to commit suicide, which are usually made at a point where the individual is not thinking clearly and calmly about the nature of his decision.  In addition, of course, this comes under the previous heading of all-too-irrevocable decisions.  Now there are practical steps which a society could take if it wanted to decrease the possibility of suicide--for example not paying social security benefits to the survivors or, as religious institutions do, not allowing persons to be buried with the same status as natural deaths.  I think we may count these as interferences with the liberty of persons to attempt suicide and the question is whether they are justifiable.

Using my argument schema the question is whether rational individuals would consent to such limitations.  I see no reason for them to consent to an absolute prohibition but I do think it is reasonable for them to agree to some kind of enforced waiting period.  Since we are all aware of the possibility of temporary states, such as great fear or depression, that are inimical to the making of well-informed and rational decisions, it would be prudent for all of us if there were some kind of institutional arrangement whereby we were restrained from making a decision which is so irreversible.  What this would be like in practice is difficult to envisage and it may be that if no practical arrangement were feasible we would have to conclude that there should be no restriction at all on this kind of action.  But we might have a "cooling off" period, in much the same way that we now require couples who file for divorce to go through a waiting period.  Or, more far-fetched, we might imagine a Suicide Board composed of a psychologist and another member picked up by applicant.  The Board would be required to meet and talk with the person proposing to take his life, though its approval would not be required.

A third class of decisions--these classes are not supposed to be disjoint--involves dangers which are either not sufficiently understood or appreciated correctly by the persons involved.  Let me illustrate, using the example of cigarette smoking, a number if possible cases.

1. A man may not know the fact--for example, smoking between one and two packs a day shortens life expectancy 6.2 years, the costs and pain of the illness caused by smoking, et cetera.
2. A man may know the facts, wish to stop smoking, but not have the requisite willpower.
3. A man may know the facts but not have them play the correct role in his calculation because, say, he discounts the danger psychologically since it is remote in time and/or inflates the attractiveness of other consequences of his decision which he regards as beneficial.

In case 1 what is called for is education, the posting of warnings, et cetera. In case two there is not theoretical problem.  We are not imposing a good on someone who rejects it.  We are simply using coercion to enable people to carry out their own goals. (Note: There obviously is a difficulty in that only a subclass of the individuals affected wish to be prevented from doing what they are doing.) In case 3 there is a sense in which we are imposing a good on someone in that given his current appraisal of the facts he doesn't wish to be restricted. But in another sense we are not imposing a good since what is being claimed--and what must be shown or at least argued for--is that an accurate accounting on his part would lead him to reject his current course of action.  Now we all know that such cases exist, that we are prone to disregarding dangers that are only possibilities, that immediate pleasures are often magnified and distorted.

If in addition the danger are severe and far-reaching, we could agree to allow the state a certain degree of power to intervene in such situations.  The difficulty is in specifying in advance, even vaguely, the class of cases in which intervention will be legitimate.  A related difficulty is that of drawing a line so that it is not the case that all ultra-hazardous activities are ruled out, for example, mountain-climbing, bull-fighting, sports-car racings, et cetera.  There are some risks--even very great ones--which a person is entitled to take with his life.

A good deal depends on the nature of the deprivation--for example, does it prevent the person from engaging in the activity completely of merely limit his participation--and how important to the nature of the activity is the absence of restriction when this is weighed against the role that the activity plays in the life of the person.  In the case of automobile seat belts, for example, the restriction is trivial in nature, interferes not at all with the use or enjoyment of the activity, and does, I am assuming, considerably reduce a high risk of serious injury.  Whereas, for example, making mountain-climbing illegal completely prevents a person from engaging in an activity which may play an important role in his life and his conception of the person he is.

In general, the easiest cases to handle are those which can be argued about in the terms which Mill thought to be so important--a concern not just for the happiness or welfare, in some broad sense. of the individual but rather a concern for the autonomy and freedom of the person.  I suggest that we would be most likely to consent to paternalism in those instances in which it preserves and enhances for the individual his ability to rationally consider and carry out his own decisions.

I have suggested in this essay a number of types of situations in which it seems plausible that rational men would agree to granting the legislative powers of a society the right to impose restrictions on what Mills calls "self-regarding" conduct.  However, rational men knowing something about the resources of ignorance, ill-will and stupidity available to the lawmakers of a society--a good case in point is the history of drug legislation in the United States--will be concerned to limit such intervention to a minimum.  I suggest in closing two principles designed to achieve this end.

In all cases of paternalistic legislation there must be a heavy and clear burden of proof placed on the authorities to demonstrate the exact nature of the harmful effects (or beneficial consequences) to be avoided (or achieved) and the probability of their occurrence.  The burden of proof here is twofold--what lawyers distinguish as the burden of going forward and the burden of persuasion.  That the authorities have the burden of going forward means that it is up to them to raise the question and bring forward evidence of the evils to be avoided.  Unlike the case of new drugs, where the manufacturer must produce some evidence that the drug has been tested and found not harmful, no citizen has to show with respect to self-regarding conduct that it is not harmful or promotes his best interest.  In addition the nature and cogency of the evidence for the harmfulness of the course of action must be set at a high level.  To paraphrase a formulation of the burden of proof for criminal proceedings--better two men ruin themselves than one man be unjustly deprived of liberty.

Finally I suggest a principle of the least restrictive alternative.  If there is an alternative way of accomplishing the desired end without restricting liberty although it may involve great expense, inconvenience, et cetera, the society must adopt it.

^* From Morality and the Law, edited by Richard A. Wasserstrom.