In discussing drug control and freedom, it still makes sense to start with John Stuart Mill's essay On Liberty (1859), for all its flaws "the clearest, most candid, most persuasive and most moving exposition of the point of view of those who desire an open and tolerant society" (Berlin 1959, p. 50). `Mill's basic principle is that the freedom of adults to live their own lives in their own way should be abridged only to protect others. No people who are "in the maturity of their faculties" and capable of "being guided to their own improvement by conviction or persuasion" (Mill 1859, chap. 1) should be forced to do anything just because it would be good for them, or forcibly prevented from doing anything just because it would be bad for them. Drug control was an important political issue at the time when On Liberty was published, and Mill had it very much in mind while writing the essay. In the name of the liberty of the Chinese opium smoker, he defended the opium trade between British India and China, which was under constant attack by reformers. He opposed laws requiring a doctor's prescription to buy certain drugs, including opiates. It goes without saying that he was against the alcohol prohibition laws passed in several U.S. states in the 1850s. In fact, he opposed most legal controls on alcohol, including penalties for drunkenness and taxes de-signed to keep consumption down (he did not object to taxation solely for revenue). Obviously he would have regarded almost all present drug laws as violations of his principle.

The libertarianism of Mill's essay, exemplified by his classic defense of freedom of speech, was regarded as radical in his own time. But his views on drug control, odd as they seem today, were largely in accord with established policies that reformers were only beginning to challenge. Today, Mill's views on free speech are considered prophetic, and his suspicion of paternalism is still attractive to many people. But only a few libertarian theorists like Thomas Szasz and Robert Nozick agree with him about drug control, and they are regarded as a radical fringe. Some study of history is ultimately needed to understand this change in attitude, but to see why there is any disagreement at all about the legitimacy and the basis of drug controls, we must first explore Mill's reasoning and the objections to it. The question is where and how Mill's principle can properly be applied, and what exceptions and qualifications have to be introduced to make it workable.

Mill was taught in the school of utilitarianism by his father James and by Jeremy Bentham, but On Liberty represents a partial break with the ideas of that school. Bentham had no objection to the exercise of paternal-istic authority. He did not regard liberty as a good in itself, but only as a political instrument that was of value wherever it helped to achieve the overriding goal of the greatest happiness for the greatest number. He did not hesitate to reconunend laws to prevent self-inflicted injury and pun-ishment for what he called "self-regarding crimes." Mill does not abandon the principle of greatest happiness — he says he will "forgo any advantage which could be derived . . . from the idea of abstract right" (Mill 1849, chap. 1), but he insists that the principle itself demands that we treat certain liberties as inviolable. (It is doubtful whether any such case can be made on purely utilitarian grounds, but that is not the issue here.)

Mill thinks that three kinds of activity should be covered by` the rule against paternalistic state coercion. Speech and opinion should be ab-solutely free; even harm to others should not be taken into account. "Combination," or joining with others to act in concert, should also be free as long as it harms no immediately identifiable innocent third parties. Finally, "tastes and pursuits," including the use of drugs and alcohol, should almost always be allowed free cultivation without government interference. Each person is the guardian of his or her own health and morals. The prohibition movement employs doctrines that "ascribe to all mankind a vested interest in each other's moral, intellectual, and even physical perfection, to be defined by each claimant according to his own standard." Of this idea he says that "there is no violation of liberty which it would not justify" (Mill 1859, chap. 4). Drug and alcohol laws, he believes, treat the people as though they were savages or children and punish all for the intemperance of a few.

Mill has to consider whether the issue of paternalism can be avoided. It might be said that no harmful act of any significance hurts only the actor; other people must be affected too. To justify a law this way is more attractive than telling protesting adults that something will be forced on them for their own good. In fact, indirect and long-term harm to others has always been the main justification given for suppressing individual drug use. The situations are familiar. Overdoses, accidents, and physical or mental illness caused by drugs may require the use of public medical resources; society may be damaged by crimes committed under the influence of drugs; drug users may neglect their families, who will require public support, or they may become unproductive and dependent on others because of chronic drug abuse.

Other devices can be used to avoid the question of paternalism. For example, to justify the helmet requirement for motorcycle riders, it has been proposed that we regard the cyclist who blithely refuses to put on a helmet as actually harming another person — the repentant cyclist after the accident. Presumably the same judgment could be made on the begin-ning drug user who gives no thought to the addict he or she may become some day. The problem is to stop the division into multiple personalities at some point before each of us is treated as a different person at each moment; without some persisting self, there can be no responsibility and no freedom.

Mill fears that indiscriminate state intervention in the name of protecting the public would reduce his antipaternalistic principle to a verbal scruple. So he insists that harm to others justifies state coercion only if it has actually occurred or is at least a definite risk. Indirect, remote, or merely possible harm is not enough. In one place he even suggests that the state should .intervenes only in cases of force, fraud, or treachery. Elsewhere he allows it more authority; for example, he says that drunkenness itself should not be punishable, but that a soldier should be punished for being drunk on duty. Still, he regards an activity as none of the state's business so long as it "neither violates any specific duty to the public, nor occasions perceptible hurt to any assignable individual" (Mill 1859, chap. 4). You may educate and persuade people not to misuse drugs (or not to use them at all), and you may punish them for actual harm to others produced by their use of drugs, but it is a clear violation of the antipaternalistic principle to prevent them by law from using drugs in the first place.

By setting narrow limits on the kinds of harm the law can talce into account, Mill avoids such difficult issues as how much productivity society has a right to demand from an individual, or how great a probability of serious harm to others constitutes a reason for state intervention. He acknowledges that some health and safety standards may be necessary to protect consumers, but otherwise he is devoted to free trade; at any rate, he believes that no one should be restrained from selling drugs, alcohol, or any other commodity to a willing buyer as long as the transaction is not fraudulent. Anything can be misused or used in excess, but the state should interfere only when the misuse hurts an innocent third party.

Mill wants to limit the numbers and kinds of acts that are described as directly harming other people. By implication, his principle also requires limits on the numbers and kinds of acts that are described as involuntary or forced. An act that is not voluntary is, in a sense, not a true act, and therefore is not protected by the rule against paternalism. Prohibiting it only prevents an internal or external force from harming the apparent actor, who is unfree and not responsible. But Mill hardly considers this question, because he takes it for granted that adults are responsible for almost all of their acts. Even the most obvious ignorance or foolishness is no justification for treating an act as if it is involuntary. He does not discuss extreme psychological and social pressures (for example, what is now vaguely called "brainwashing"), but in any case it is questionable to describe that way the conditions that produce most drug and alcohol abuse. He does admit that people might need restraint to prevent them from exposing themselves to danger while "in some state of excitement or absorption incompatible with the full use of the reflecting faculty" (Mill 1859, chap. 5), but he understands this to be a special, rare, and temporary condition; today we tend to treat it as a persistent and recurrent one. For example, the legal theorist H. L. A. Hart, a successor in Mill's liberal tradition, criticizes him for using his principle to disallow restraints on self-inflicted injury caused by inadequate reflection, transitory impulses, weak will, and other subtle pressures (Hart 1963). But few actions are entirely free of all these defects, so taking them into account might in effect nullify the prohibition on paternalism. That is what drug laws, like consumer protection regulations, are designed to do.

Certainly we are much less inclined than Mill to treat drug use as the free, rational act of an autonomous person. It is usually considered a product of ignorance, impulsiveness, or, worst of all, addiction — conditions in which the drug user's freedom becomes a minor concern. For example, almost every heroin addict is introduced to the drug by friends w'ho use it but are not addicted and never become addicted. But we do not want to believe that anyone takes this drug out of free will and with full knowledge: The user must be mentally disturbed, a child victim of evil adults, or, finally, a slave to addiction. So we have developed the myth of the pusher who lures innocent children with free samples. An addict who began to use heroin as a child can never have been free to choose whether or not to use it.

The easiest way to limit Mill's principle, then, is to accept it but at the same time to insist on taking into account many external causes and consequences that Mill himself would have considered politically and legally irrelevant. Acts that Mill would have regarded as free are treated as in some way socially or psychologically compelled; effects on other people that Mill would have considered too remote, indirect, secondary, or uncertain to require government intervention are treated as good reasons to pass preventive laws. Contemporary drug laws are readily justified in this way without appealing to paternalism in the strict sense.

Maybe it is true that the whole concept of paternalism is otiose in this case, because drug use is intrinsically less free than most other acts, or because the extent and severity of the drug problem give us good reasons to take into account consequences of an individual act that would other-wise be considered too remote or uncertain. We will eventually explore these questions. Meanwhile, let us assume what seems intuitively plausible to Mill — that some laws, including many drug laws, are truly paternalistic, directed mainly at self-injury.

These laws can still be defended, but only if Mill's two strong arguments against paternalism are rejected. The first one is that people in general are the best judges of their own interests. The opinion of others about what would benefit me is likely to be wrong when it conflicts with my own opinion; when authority tries to force me to do what is good for me, it ifi likely to interfere in the wrong place, at the wrong time, in the wrong way. Even when someone seems to be mistaken, it is best to educate and persuade rather than to use coercion, and resorting to force discredits those better means. Mill's second objection to paternalistic coercion is that it prevents people from making "experiments of living" (Mill 1859, chap. 3) that encourage the realization of individual potentialities, development of human faculties, variety of character, and richness of experience. By prejudging the value of the individual's desires and ends, paternalism imposes a conformity that limits human possibilities.

Almost no one rejects these arguments outright, but many people insist that the exceptions are much more numerous and, above all, more easily identified than Mill thinks. They believe that they can clearly define situations in which some external authority is able to judge a person's interests better than he or she can. They also claim to identify experiments in living that do not encourage individuality, diversity, and the fulfillment of human capacities. In some of these situations, education and persuasion alone are inadequate, because there is not enough time for them to work before serious harm occurs or simply not enough chance that they will ever work. Recreational drug use is usually put in that class.

Childhood is the most important situation in which we universally recognize some paternalistic discipline as necessary. Children are not assumed to be the best judges of their own interests, and restricting the experiments in living they are allowed to undertake is thought to promote rather than hinder their self-realization and the development of their individuality. The very word "paternalism" implies that the paradigm of coercing a person for his or her own good is the biologically natural authority of adults over children. That is why Mill insists that his principle applies only to people in the maturity of their faculties. Most of the reasons given for paternalistic coercion of adults are related to this model of childhood. They all assume that beneficiaries of paternal discipline are in some way incomplete — not, or not yet, fully developed human beings who must be granted the right to determine the course of their own lives. Even a writer who calls himself an anarchist can criticize "absolute tolerance" by comparing it to permissive child rearing (Wolff 1965). One cotmnon assumption is that whole societies and peoples can be historically immature, as a child is biologically itmnature; they require paternal gov-ernment to realize their full historical potential. For this reason, Mill is willing to allow paternalism in the societies he calls "savage" and "bar-barian"; in fact, one of his objections to alcohol control laws is that they treat the English laborer like a savage. Mill assumes a special status for nineteenth-century industrial society, but there is no need to follow him Marxists, for example, insist that every class society is historically immature. We will cease to be barbarians only when humanity comes to its self-realization under communism. Until then, at least, Mill's worries about paternalism are beside the point.

But liberals generally refuse to treat every existing impedection in individuals and societies as analogous to the biological immaturity that makes a parent's authority necessary. They require another principle to decide when individuals are so imperfect or incompetent that the state must protect them from themselves. One idea is that the state should be allowed to intervene whenever the situation is one in which a reasibnable person would choose to be protected from the consequences of doing what he or she wants. The unreasonable person in the actual situation — ignorant, incompetent, foolish, impulsive, driven by forces beyond in-dividual control — is placed under the authority of a hypothetical reasonable self whose agent is the state insofar as its laws are just and sensible. In effect, the coercive external authority is referred back to an individual taking prudent precautions for self-protection. And, as Mill insists, persons are usually the best judge of their own interests.

Gerald Dworkin has represented this situation in its simplest form by reference to the myth of Ulysses and the Sirens. Ulysses wants to hear the Sirens sing, but he fears being lured to death by drowning. So he has himself lashed to the mast of his ship and orders his men (who will be protected from the Sirens by wax in their ears) to ignore the anguished demands for release that he knows he will make when he hears the seductive and deadly song. When the sailors refuse to untie Ulysses, it could be said that they are exercising paternalistic authority over him; and yet, from another, more obvious point of view, it is they who are recognizing his authority. No moral or political problem arises here, and it is certain that when he comes to his senses Ulysses will approve of the sailors' actions (Dworkin 1971). The alcoholic writer Jack London is said to have proclaimed that he favored women's suffrage because the women would enact prohibition laws and save him from himself. This anecdote suggests a modern political version of the Ulysses story.

Unfortunately, in politics the command of the reasonable self is rarely so explicit, and a person who has not explicitly asked for the restraints is not nearly so certain to approve of them afterward. For that matter, in practice it is not easy to tell which self is the reasonable one. John Rawls touches on the problem with his theory of the original position in which individuals meet to frame a social contract behind a veil of ignorance. Each fictive person in this imaginary situation is conceived as fully devoted to rational self-interest but completely ignorant of his or her personal characteristics and social circumstances in any real world. Rawls believes that partrnalistic coercion or restraint is appropriate whenever someone in the original position would judge it necessary to ensure against the possibility of undeveloped powers or self-destructive impulses. In a way, this person is the solicitous parent of the empirical self — a child with whom he or she is truly one flesh. No one could be a better judge of the interests 6f this imperfect real person or a better guide to that person's self-development and self-realization, so both of Mill's arguments against paternalism are eliminated. Judgment on what constitutes self-destruc-tiveness or undeveloped powers is left first to the individual's rational, settled preferences. If we cannot say what these preferences are in a given situation, we must introduce a theory of "primary goods" — a notion of what is good for people in general (Rawls 1971, pp. 248 ff ).

This takes us a long way from Ulysses' command and an even longer way from Mill's principle. Mill mistrusts paternalism because he believes that no conceptions of what is desirable for human beings are so fixed and certain that they need only be applied by a wise authority. But Rawls's abstract person in the original position, the creator of the social contract, is made out to be just that kind of authority on the good. The proper exercise of liberty in effect becomes identified with choosing rightly or rationally by a standard that has nothing to do with the individual's actual desires. 'The effect is a broad license for paternalistic restraint, ostensibly to promote genuine freedom.

Suppose that Ulysses forgets to order his men to ignore later commands given under the influence of the Sirens. Still, few would object if they denied his wishes for his own good. By refusing to untie him, they would be obeying the orders that an ideal Ulysses would have given, and the actual Ulysses will thank them for it later. Rawls might call this deference to his rational settled preferences, but he also might call it an application of the theory of primary goods. In this case, life is the primary good that would be chosen by the rational person in the original position. Even if the sailors do not know their captain's rational settled preferences — even if they imagine that later he might tell them that he would rather have died a blissful death — they should rely on the notion of primary goods and treat him as though they were obeying the command of a man in the original position.

A more practical example is legislation setting maximum hours and minimum wages for labor. In effect, the individual worker's freedom of contract, such as it is, is limited for the collective benefit of all workers. Workers actually fight for this ostensible limitation on their individual freedom, and they certainly do not regard it as paternalistic. Nor does Mill; he insists that such laws actually grant rather than deny power to the people affected by them. But this is ostensibly a case of paternalism, since the law is being used to protect workers against temptation. A worker who votes for a law or signs a union contract setting maximum hours and minimum wages is in somewhat the same position as Ulysses commanding that certain wishes he might later express should be denied. But these laws and contracts are also enforced against (or in favor of ) workers who have not agreed to them and may even object to them. Ulysses neglecting to give his command was not in the same situation. There the sailors could assume an oversight, because refusing to untie him served not only his own rational preferences but also a primary good recognized by everyone. Mill would certainly have judged Ulyssts to be in a state "incompatible with the use of the reflecting faculty," but even here some extreme libertarian might defend letting him go. In most cases it is less clear what the person under restraint would prefer if he or she were reasonable, and in many cases it is not obvious to everyone that a primary good is being served. For workers to believe that wage and hour limitations do them more harm than good is not obviously irrational, although it may be wrong. So a few individualist libertarians would allow maximum hour and minimum wage restraints to be applied not to workers in general but only to those who have actually contracted for them.

The larger issue here is which self-injurious actions are compatible with rationality and therefore with freedom. If free acts, however wrong or mistaken, are in some sense products of a fully developed rational capacity, it is hard to see how Mill can dispense with some conception of human nature or the meaning of human life that gives content to the idea of rationality. He actually alludes to this in the epigraph of his essay, where he quotes Wilhelm von Humboldt as saying that the aim of government is human development in its richest diversity. Mill objects to paternalism partly because it assumes the existence of primary goods or rational preferences so obvious that they can be imposed on people against their will. But to speak of rich human development implies some conception of fulfillment appropriate to human beings and therefore some idea of rationality as a rule directing action toward such a fulfillment. By that standard some acts must be irrational and, if there is no freedom without rationality, possibly unfree. And maybe such acts should be prevented by paternalistic coercion.

Mill himself hardly considers the matter further after malcing exceptions for children (and "savages") and for acts performed in a state of "excitement or absorption." But promoting the highest development of human powers could easily require much more work by government. The tension is inherent in Mill's conception of the individual; he is uncertain which of our characteristics demand that freedom should be defended against pa-ternalistic coercion. He is concerned about developing and exercising the capacity to conceive our own ends, shape the meaning and course of our own lives, and maintain our own conceptions of happiness, so he emphasizes the value of the individuality and diversity that result from such choices. But it is easy to think of acts (including some kinds of drug abuse) that are free by Mill's definition, yet actually seem to destroy the capacity to shape our lives and therefore reduce individuality and diversity. The shaping of my life should be guided by what is significant for me, but I am not always the best judge of that. And even when I do know what is really important for me, I may be the victim of other desires that prevent me from attaining it — desires I wish I did not have. Alcohol and drug dependence are the paradigms. A life dominated by impulse, drift, or restricting habits is not free. Mill himself actually writes in Logic that "none but a person of confirmed virtue is completely free" (Himmelfarb 1974, p. 108). This conception of freedom is obviously not the same as the political one of On Liberty.

The trouble is that to justify paternalism by distinguishing meaningfully free acts from those produced by impulse, drift, and compulsive habit would be to smuggle in the idea of obvious primary goods and universal rational preferences that he has already rejected by denying that there is any fixed human nature capable of serving as a political standard. From another point of view, the distinction implies that a more or less balanced, fully developed character is required for politically protected free action.

But once we introduce that criterion, it is hard to know where to stop short of identifying free acts with rational and virtuous acts as defined by a standard independent of the individual's will. And Mill fears the political implications of doing that.

The requirement of wholeness, integrity, consistency, and ultimately rationality as conditions for freedom can be extended beyond the individual. No one is self-sufficient; no one is even himself or herself outside of a context created by other people in families and societies. Mill's distinction between harm to oneself and harm to others becomes unimportant in this perspective. Any injury done anywhere affects both individual interests and the common good at once, and the freedom of a whole person is also the good of a whole society. These ideas are worked out most clearly in the theory of positive liberty. Mill's contemporary, T. H. Green, used the term as a contrast to negative liberty, permission to do what you want, which is central to Mill's conception of political freedom. The distinction has been analyzed by Isaiah Berlin (Berlin 1969). Advocates of positive liberty insist that a free individual must be a whole person, and no genuinely whole person can exist outside of a free society, that is, one which allows the full development and self-realization of each person. As long as social arrangements are imperfect, all individuals are unrealized, incomplete, and to that extent, at least, like childfén. In the terminology of Hegelian idealism, the empirical will, which exercises the dubious freedom Mill defends, is distinct from the real or essential will , which is in total harmony with reason and can have full effect only within a society that is also totally reasonable. In Marxist terminology, people in a class society are alienated, estranged from their true hatures. They are in a socially transitional stage; alienation will end only when the human species comes to full maturity under communism. In either view, freedom is best conceived not as the exercise of an imperfect individual's will but as the product of a good life in a reasonable society. What Mill regards as paternalism may be one of the ways to attain that freedom.

The idea of positive liberty goes back in Western thought at least to the remark of Socrates that no one willingly errs. Montesquieu characterized this concept succinctly in his statement that liberty consists of being able to do what you ought to want to do. In this tradition, persons who consider themselves free may be slaves to impulses that destroy the capacity for meaningful, rational choice. All action opposed to the common good is unfree in the deepest sense, and it may be necessary to coerce people for an end they are too depraved or foolish to recognize — as Rousseau put it, to force them to be free. Green (1895) defines natural rights as the powers individuals need to attain the ends human society is meant to realize. He believes that many restraints called paternalistic by Mill, such as laws restricting the use of alcohol and drugs, actually create new rights by preventing behavior that interferes with the possibility of free action contributing to the public good. Benevolent restraint nullifies a hidden coercion that destroys human powers and therefore violates natural rights.

The central question about all paternalistic authority is who will guard the guardians, or, to put it more abstractly, how the authority is to be appointed and what justifies its rule. Where the beneficiary is a child, as we saw, the answer is easy: The authority is appointed by nature and its justification is maturity. A problem arises only where adults are involved. Marxists often vest authority in those social classes or groups that are historically mature and therefore understand the present potentialities of humanity. Rawls believes that just laws, including paternalistic restraints, derive their authority from the refined rationality of the person in the original. position; we need only imagine ourselves to be divining the thoughts and wishes of that person, which are those of an ideal parent who is also, somehow, ourself. Superior hereditary capacity and a vision of the idea of the good give the guardians of Plato's Republic the right to rule.

But there are always doubts about what group is historically mature, what the person in the original position would think, who has been granted a vision of the good. And as long as this is so, paternalistic authority may be under suspicion as a disguise for tyranny and a danger rather than an aid to the development of human capacities. Kant writes of the need to liberate humanity from its self-imposed tutelage and says that paternalism is the greatest despotism imaginable. Von Humboldt says that the incapacity for freedom can arise only from a want of the moral and intellectual power supplied by freedom itself (Chomsky 1973, p. 398). Many advocates of positive liberty also see the problem. Green admits that some kinds of external restraint reduce "the capacity for spontaneous action regulated by a conception of a common good" (Green 1895, para. 209) and thus retard the development of freedom; these are the restraints he is inclined to regard as paternalistic. In fact, he believes that very few acts are so necessary or so dangerous to the public that we must enforce or prevent them with threats of punishment. He admits, in effect, that negative liberty is necessary to develop the capacity for positive liberty — a point not far from Mill's justification of his antipaternalistic principle as a way to develop human powers.

Some compromise is necessary. Everyone wants to put limits on the state. The term "paternalism" is suspect even to Green and others who approve of laws that Mill would certainly have called paternalistic. Rawls would say that we must be left to make our own decisions about some things, either because the rational person of the original position would have no opinion about them or because he would believe that we should be free to choose. The Marxist has to consider whether some kinds of benevolent restraint are incompatible with the historical level of social forces and therefore unacceptable to their beneficiaries. Green admits that not every power whose exercise would ideally be desirable can be claimed as a right in a given society at a given time; the claim must also be acceptable to the public. Even supposing it desirable that no one should ever drink alcohol, the state would have no right (as well as no capacity) to enforce a prohibition law if the populace was not willing to tolerate it.

The idea of positive liberty in its purest form actually tends toward anarchy rather than any sort of state control. It suggests a vision of a world in which my freedom is always fully compatible with yours, total self-mastery is joined with complete absence of overt coercion, and impulse coincides with deepest reflection in a perfect harmony of will and reason. At some ideal point humanity reaches adulthood, its compulsory education ends, and there is no more need for coercive law or a state. Marxists say that the state will wither away when the condition for the free development of each is the free development of all. Green's political theory has similar implications: "For a positive and detailed criterion of just punishment, we must wait till a system of rights has been established in which the claims of all men . . . are perfectly harmonized. . . . And this is perhaps equivalent to saying that no complete criterion of just punishment can be arrived at till punishment is no longer necessary; for the state of things supposed could scarcely be realized without bringing with it an extinction of the tendencies which state-punishment is needed to suppress" (Green 1895, para. 189). On the assumption described (and rejected) b,y Isaiah Berlin, that the rational ends of our true natures must coincide (Berlin 1969, p. 154), there is no right to negative liberty in existing societies, but no need for any coercive state once our true natures have been realized. Mill, in contrast, does not believe that even the highest and most harmonious development of human powers could ever become so perfect that both coercion and formally prescribed liberties would be unnecessary.

Theorists of positive liberty seem to be justifying a strong state authority under some conditions and an ideal anarchy under others. But in either case they treat the distinction between harm to self and harm to others as unclear and difficult to apply in practice. The result is the same as if they avoided the issue of paternalism entirely by finding, in every act the state proposes to restrain, causes beyond the actor's control or consequences affecting other people. So long as no one even has a self outside of a social context, any harm that a person does must have social as well as personal causes and effects. Theories of positive liberty therefore have the same effect as insistence on external causes and consequences.

Whether they speak of positive liberty or external harm, the critics of NMI's principle are making an important point. In modern societies many laws, including some of which almost everyone approves, can be regarded as violating Mill's principle on Mill's interpretation of it. As Green points out, the principle is much too broad even to fit our commonsense notions of what is paternalistic. The laws that govern building inspection, practicing medicine without a license, minimum wages, consent as a defense in assault charges, the dispensing of prescription drugs, seat belts in cars, gambling, obscenity, prostitution, swimming in public pools without life-guards, laetrile, food additives, dueling, suicide, and selling oneself into slavery have little in common. They cover very different areas of social life, they were established in very different circumstances, and they are enforced in very different ways. Some deal with what we ordinarily call consensual crimes, and others do not; some would be regarded as paternalistic by most people, and others would not. It is true that all of these laws involve or might involve violations, modifications, or qualifications of Mill's principle, but that does not tell us much about them, and it gives little guidance to legislators or the public. Knowing how to regulate minimum wages, dueling, or the use of swimming pools hardly helps in handling the problem of recreational drug use. Analogies that may be interesting to political theorists seem to have little practical use. It is not surprising that most people prefer to avoid Mill's formulation of the issues. Certainly legislators rarely consider it on his terms; they are no more disinclined to enforce, punish, or prevent self-regarding acts than Bentham was.

Yet, as we saw, even among writers who reject most of Mill's applications of his principle and most of the theory behind it, a feeling persists that there is something worth preserving, some inner core of negative liberty that should be inviolable. Unfortunately, it is hard to define that inner core in a principled way that is neither so broad and abstract that it takes in almost everything we do, nor so limited and concrete that it amounts to nothing more than a list of policy recommendations for specific situations. For example, Ronald Dworkin proposes a distinction between the "force" of Mill's principle, which he says is great, and its "range," which he considers quite limited. Liberty as "independence" is what Mill endorses, not liberty as "licence" (Dworkin 1970, pp. 261ff.). But unless this distinction is the same as the one between positive and negative liberty, it is merely rhetorical. Dworlcin seems to be using it to justify ad hoc whatever policies he considers desirable. The "force" of the argument against state interference with sexual acts between consenting adults is said to be enormously powerful because sex comes within the proper "range" of the principle; but outside that range, in the territory of drug use or consumer protection, the principle may have no force at all. This is a statement of preference, not an argument, and it is not Mill's prefer-ence — he defends the sale of potentially deleterious commodities.

To resolve the issue, we need to know not only whether laws on drugs, sex, or any other subject can be justified on some grounds, paternalistic or nonpaternalistic, but also why paternalistic or preventive legislation is socially distributed the way it is. Why, in societies where so many actions dangerous to the actor and to others are permitted and sometimes en-couraged, do we impose such strong restrictions on drug use? Why are these laws often controversial, and always difficult and expensive to enforce?

Utilitarian calculation might seem capable of proving that a certain system of control does more good than harm when applied to drug use but would do more harm than good in some other area of social life. But when an issue is at all complicated, the utilitarian calculus simply does not supply the necessary distinctions; comparing different kinds of benefit and harm on a single scale is too difficult. It is important to' mention, incidentally, that in actual policy discussions about drugs, we rarely concede that there are any benefits to be had from using them. As the Canadian Corrunission of Inquiry into the Non-Medical Use of Drugs said in its 1973 report, there is "no clearly established framework" for assessing the good effects of drugs (Final Report 1973, p. 50). Legislators show no interest in whether or when recreational drug use does any good, and little interest in comparing the harm done by drugs to other kinds of harm. So in practice, the only utilitarian issue is whether the reduction in harm caused by the drugs is worth the cost of enforcing a drug law. And even that issue is rarely given much serious attention.

Analogies do help to explain the controls on recreational drug use, but not just anything that seems paternalistic will do. The present system of drug regulation can be analyzed using three analogies: rules relating to dangerous sports and hobbies; consumer protection laws, including public health and the control of medical practices and technologies; and the regulation of public morality or public order. Twentieth-century govern-ments have been forced to recognize, implicitly or explicitly, the power of these analogies, which in their encounter have produced the history of modern drug laws and the form taken by our modern drug problem.

In practice, the analogies are used in combination to justify severe social disapproval and legal restraints that we would probably not tolerate for equivalent dangers (if it were possible to speak of them) produced by acts other than drug use, while at the same time supplying apparently inconsistent exceptions for two especially dangerous drugs, alcohol and tobacco. The current drug control arrangements and their justifying analogies have the weight of twentieth-century history behind them, and they are similar everywhere in the world. They are the accepted and perhaps inevitable response, in our time, to a need for classification and control of a complicated and ambiguous social situation. Whether we want to protest against this system, modify it, or defend it, we need to become more aware of just what we are doing and describe it more candidly.

Consider a typical article on the dangers of what is colloquially known as "getting high." The alarming interest in this dangerous behavior is pointed out, its social roots are examined, the epidemiology of the habit is analyzed, and the motives of the young people who take the trips are discussed. Some are depressed, others bored; some want escape, adventure, or a way of testing themselves in the search for identity. A few have personality disorders, and others suffer from feelings of inadequacy. The roles of parents, peers, and social conditions are considered. The author notes the euphoria often produced by the experience and describes the process of going up to a peak and coining down. The sometimes remarkably compleZand expensive paraphernalia used and the proliferation of shops displaying and selling it without interference from the law are mentioned and deplored. The abnormal mental and physical states produced by the habit are analyzed, and the danger of serious accidents, even death, is emphasized. The question of psychological dependence is discussed. The article ends with a call for more effective law enforcement and a re-examination of the moral and spiritual condition of the society that has produced the epideinic.

It sounds unremarkable; the ideas and vocabulary are familiar to the point of boredom. But a few years ago, when a satirical piece with this theme was published, its title was "Alpinism: The Social, Scientific and Treatment Aspects of Getting High and Its Prohibition" (Phallow 1978). The point, of course, is that much of the descriptive language used in writing about drugs applies to mountain climbing as well; only the customary denunciations, anxious soul searching, and earnest recommendations sound absurd when transferred. For mountain climbing you can also substitute the still more dangerous sport of flying small planes (Lindsey 1980) or even, say, high school football, which produces several dozen cases a year of permanent quadriplegia (Torg et al. 1975). It would be silly to suggest that drug use is just like these other pursuits and should be subject to similar rules, but the analogy is not entirely irrelevant. As the satire implies, we tend to ignore similarities between drug use and other voluntary activities that may injure the participant. Obviously we do not want to find ourselves comparing the effects of marihuana on high school students with the effects of football.

Mill supplies the justification for not prohibiting dangerous sports; it preserves liberty by protecting individual tastes and pursuits from pater-nalistic coercion. The same reasons can be given for not prohibiting potentially unhealthy eating habits. But official policy rejects the notion that a drug trip might resemble an airplane trip or a drug habit might resemble a habit of eating candy. One possible reason is that drug use is not considered a free act in the same sense as eating candy or flying small planes; another is that, unlike these other activities, drug use has little or no value that justifies the risk. In practice, the two ideas reinforce each other, because of the connection between freedom and rationality. If we decide that something has no value and its pursuit is therefore irrational, we look for ways to show that people who want it are not making a truly free choice.

Mountain climbing is dangerous, but its pleasures are virtuous ones; it provides an opportunity for physical exercise, adventure, achievement, and the enjoyment of natural beauty. In making public policy, the illness caused by enjoyable but unhealthy eating habits or the accidents suffered by careless mountain climbers and motorcycle riders are balanced against the satisfaction of human needs and desires that these activities are assumed to provide. But drug use for anything but the treatment of disease is not regarded as satisfying a legitimate need or desire. There is some underground lore about the alleged beneficial effects of illicit drugs, and there is also a more respectable lore about the virtues of alcohol. But they tend to be nervously jocular or embarrassed, and have little effect on public policy. Permitting drug use is often defended in the name of individual freedom, but rarely on the ground that there is any good in it.

As Mill apparently fails to understand, in this case circumstances can make talk of individual liberty trivial or even hypocritical. In a complex industrial society, where far more needs to be known than any single person can learn, there are some things about which the average person is simply incapable of making an informed and rational choice. To use the jargon of neoliberal economics, the cost of information processing is too high. In less scientistic language, everyone is in a sense limited and tmderdeveloped for this purpose, and therefore needs the kind of protection Mill mistrusted as paternalistic (we may also call it, more aggressively, "consumers' rights"). Outlawing marihuana smoking is not like oudawing motorcycle riding, a taste or pursuit to which people have a right, however dangerous it may be to themselves and others. Instead it is seen as something like outlawing the sale and use of a motorcycle with a defective part. People might be willing to buy that motorcycle, but only out of ignorance, carelessness, or impulsive folly. Since it has practically no value, society does not have to respect the freedom (if it can be called that) of those who choose to do it. Another example is the regulation ensuring that electric chainsaws cannot be sold without shields to protect the user. Someone might want to buy a chainsaw without a guard, because it is cheaper or less clumsy to handle, but we do not respect that wish. Using a chainsaw without a shield may not be more dangerous than alpine climbing, but we also assume that it has little value. These analogies could be carried further. Drugs with no accepted medical uses are treated likedefective motorcycles; they are pleasure vehicles that have a deadly flaw. Only in this case, the flaw is such an integral part of the design that the vehicle itself must be banned. Drugs with medical uses are more like the chainsaw — acceptable as a tool for certain purposes, but only with safeguards that cannot be left to the individual user to supply.

Certainly drugs are potentially dangerous instruments that can be ig-norantly misused and sometimes produce serious ill effects. But it is hard to believe that the risk of, say, marihuana smoking is so obviously great, the benefits so obviously negligible, and the consumer's ignorance so substantial that the drug has to be treated as something from which all consumers must be protected. Consumer protection laws imply that some-times we cannot rely on the commonsense rule that greater dangers are also much more obvious and therefore more likely to be avoided. People are not scientific calculators of risk, and they tend to underestimate the probability of unlikely events. With a product like the chainsaw, the danger might seem too remote and contingent to outweigh the immediate inconvenience of a safety device if there were no special regulation. But presumably the more you thinlc about it, the more you are likely to demand a chainsaw with a shield. The situation is different with drugs. Studies show that the less people know about the effects of recreational drugs, the more dangerous they consider the drugs to be (Glaser and Snow 1969; Swisher 1971); this has even been used as an argument against drug education in schools. If people who know more about the facts are wrong in their attitudes, which is quite possible, the mistake is obviously not caused by the kind of ignorance or preoccupation usually said to justify consumer protection laws.

Another peculiarity of drug laws as a form of consumer protection can be illustrated by comparing them with pollution control laws. Here we have to balance two goods, industrial productivity and a clean environment. We try to estimate how much loss of productivity we can tolerate for the sake of a given degree of improvement in air or water quality. If we followed the same policy in regulating drugs, we would try to estimate how much loss in the benefits of recreational drug use we should tolerate for the sake of a given reduction in their ill effects. But in fact the issue does not even arise, since benefits from recreational drug use are not conceded to exist for policymaking purposes.

An even more interesting analogy is the laws on seatbelts in cars. It took a long political struggle to get them installed, and now most people will not use them. But it is still very doubtful, in this country, that buckling of seatbelts will be made compulsory. Most legislators who regard driving without a seatbelt as a right would reject as absurd the notion of a right to smoke marihuana. And yet the argument against individual freedom of choice seems at first glance much better in the seatbelt case. The risk of driving without a seatbelt is overwhelmingly greater than any benefit, a fact not nearly so obvious in the case of marihuana; and driving without a seatbelt is not an active taste or pursuit like marihuana smoking, so it is more likely to be done automatically, without conscious Choice or genuine thought — the kind of situation in which people need protection most and the threat to their freedom is least.

Driving without a seatbelt resembles drug use because it may cause accidental harm. Looking at television is sometimes said to resemble drug use in another way: It is said to be an "addictive," psycholokically pernicious, and socially debilitating practice, especially dangerous to children, which can seriously damage the quality of life. But we have no laws defining a permissible amount of television viewing, just as we have no laws requiring the use of seatbelts. It is true that these matters are not left to individual choice everywhere. In many countries drivers are fined for not wearing seatbelts; in some places television viewing is restricted by law. What is interesting is that drug use, except for alcohol, is left to individual choice nowhere. Drug laws are not typical protective laws, but a special set of regulations for a very special case.

Another anomaly makes this situation clearer. If it ever became necessary for the government to use vast amounts of money and personnel to curb an organized illicit traffic in chainsaws without shields, or most other commodities forbidden by consumer protection laws, the law would probably be repealed. If people wanted the commodities so much, we might conclude that they have a legitimate interest and value strong enough to outweigh any argument for prohibition. In other words, we would handle the problem as we handle mountain climbing, hang-gliding, or motorcycle racing: We would treat it as a matter of preferred tastes and activities (however questionable) rather than consumer error. But it is considered normal and necessary for the police to devote a large part of their resources to combatting the drug traffic, and the size of the problem is not regarded as a reason to change our way of dealing with it. Besides, every activity connected with any of the banned drugs is a crime, including simple possession. Possessing some of these drugs is still a felony in many states, and the sale of a few grams is often subject to the same punishment as rape, armed robbery, and second-degree murder. Obviously more powerful feelings are at work here than those that produce the average consumer safety law.

If the comparison with seatbelts or chainsaws sounds strained, that is because psychoactive drug control is usually regarded as something more than an ordinary health and safety measure. The regulations for drugs of pleasure, drugs of abuse, or "controlled substances" are not the same as the controls on other drugs. The two systems have different historical origins (regulation of so-called narcotics started much earlier) and remain different today, despite some convergence. The controls on pleasure- and performance-enhancing drugs not only protect consumers but also contain what is believed to be a threat to the social fabric and the moral order. This is „the third analogy popularly used in thinlcing about drugs: They belong to the group of problems that includes prostitution, pornography, and gambling. Iran since the revolution of 1979 supplies another example, more interesting because more alien. A police campaign has been instituted there against alcohol, drugs, and Western popular music, which are all regarded as parts of a single social problem. The music has to be outlawed because it is "addictive," causes disorderly conduct and sexual excess, and ultimately produces undesirable moral and social changes. In other words, the Iranians think of music the way we think of drugs (and a few other vices). We do not believe that governments have a right to issue decrees on the forms and uses of music, and most of us do not regard musical listening habits as a moral issue. What makes us so sure that legal restraints on music are not just as sensible, or just as pointless, as legal restraints on drug use?

'The defense of public morality can be used as a justification for ignoring Mill's principle in two familiar ways: Either the victims of gambling, drug, and sexual habits are not free, or their behavior by its very nature harms others. But the legislation of morality has been much more disputed in modern liberal societies than consumer protection laws, usually because of doubts about the notion of morality involved and the assumed connection between the disapproved behavior and various social ills. A public debate on the subject took place in England in the early 1960s, when the jurist Lord Devlin defended laws against homosexual acts. He argued that when a society decides whether to make an act a criminal offense, it must consider the effect on the common morality as well as harm done to a victim. A society, he said, is a community of ideas, and acts offensive to deep common convictions disintegrate the bonds that enable people to live with one another. For example, the principle barring consent or forgiveness as a defense to an assault charge ensures that even if the victim acquiesces in violence, society will enforce its own idea of what is intolerable among human beings living together in a community. When the man on the street, or, as English lawyers say, "the man on the Clapham omnibus," feels strong enough intolerance and disgust, the law must ratify his feelings for the sake of social cohesion (Devlin 1965).

Several writers challenged Devlin's argument. The legal theorist H. L. A. Hart argued that feelings of intolerance and disgust do not constitute a moral conviction (Hart 1963), and Ronald Dworkin later expafided this argument into a distinction between "sociological" and "discriminatory" senses of the word "morality." In the first sense, morality includes anything that any sufficiently large social group considers to be part of its system of ethical beliefs. In the second sense, a conviction is not moral unless the person who holds it is able to give reasons for it — reasdns not contaminated by prejudice (for example, racism), personal emotion, false factual beliefs, or rationalization, and not dependent solely on the beliefs of other people; for example, "Everyone knows that homosexuality is a sin" is not a discriminating moral statement. A certain degree of sincerity, consistency, and lack of arbitrariness is also necessary; to say that the wrongness of homosexuality is self-evident, like the wrongness of cruelty, would be to fail this test. Dworkin says that Devlin wants the law to ratify morality in the sociological sense, although in fact it should be concerned only with morality in the discriminatory sense. A consensus of prejudices and passions that presents itself as morality deserves no legal respect, and by ignoring it, a legislator or judge vindicates the true implicit morality of the community (Dworkin 1971, pp. 248-55).

Hart adds that Devlin assumes a degree of social solidarity not common anywhere at any time and especially unlikely in modern society. He identifies Devlin's idea with what the nineteenth-century judge Leslie Stephen called the "denunciatory" function of punishment. But without the backing of an overwhelming moral majority, denunciation by official punishment becomes merely an idiosyncratic and excessive expression of anger and disgust. We have no single universally accepted code of sexual conduct, only a number of mutually tolerant sexual moralities.

Hart also points out that in his defense of legal moralism, Devlin effectively defines a community as a set of shared moral assumptions, and therefore allows no way to distinguish between a breakdown in the social order and the emergence of a new moral consensus. He implies that any unpunished act defying the moral consensus in any area of social life is a potential threat not just to a particular custom but to the whole life of the community. On this assumption the fabric of society is woven so that a break anywhere causes it to unravel everywhere. Moral change becomes either impossible or catastrophic. Devlin responds that not every deviation from the accepted code is dangerous enough to be made a crime, but he gives no rule for deciding which ones are — or, rather, no rule except the lawmaker's estimate of the intensity of popular feeling, the intolerance and disgust felt by the man on the Clapham omnibus (Devlin 1971). Although Mill does not discuss legal moralism explicitly, he anticipates this response when he criticizes those who "consider as an injury to themselves any conduct which they have a distaste for" (Mill 1859, chap. 4). So Mill, Hart, and Dworkin all agree that Devlin tries to introduce legal coercion where it is not desirable, but only Mill would regard drug laws as illegitimate coercion of that kind.

Devlin's argument derives some of its plausibility from being an extension of a special case: the need to preserve public order and decency against offensive public behavior. Laws against the open display of obscene material, sexual acts in public places, or public intoxication might be justified because the affront to the sensibilities of the majority is immediate, direct, and unavoidable. In this case it may make some sense to regard producing disgust or annoyance as a form of injury. The laws defended by Devlin treat private behavior as though it were a public nuisance, a stink in the public nostrils. Devlin's legal moralism implies that if the very thought of someone looking at pornography or performing strange sex acts or smoking marihuana offends people enough, it should be outlawed just as if they were being unwillingly confronted with it on the street — having the smoke blown into their faces, so to speak. This is an old idea. Brothels used to be called "disorderly houses," a term that assimilated breach of the moral order to breach of the public peace. But the modern tendency has been to insist more and more on the difference. For example, in New England in the seventeenth century, courts interpreted common law doctrines on drunkenness to make intoxication in private a crime, but later reinterpretations established that the law's concern was not sin but breach of the peace. Devlin's implicit blend of moral disorder with offensive public behavior represents a reaction against that historical trend.

Another way in which legal moralism can be injected into criminal law is the notion of alarm and fear as types of harm that must be taken into account in calculating the utilitarian balance. Bentham feels no reverence for tradition, and he certainly does not credit popular feelings of intoler-ance and disgust with special moral weight. But for him, bad feelings, especially if they are strong enough to be described as "alarm," are bad things; and to the extent that any behavior produces general alarm, it works against the greatest good of the greatest number. This is only one consideration among many in deciding what to treat as an offense or crime, but it could be important. General alarm is related to the concrete fear provoked by a robbery or assault in somewhat the same way that the offensiveness of private drug use and unusual sexual behavior is related to the offensiveness of public intoxication or obscene gestures in public. If marihuana use, for whatever reason, provokes this kind of public alarm, it may have to be treated like robbery and assault in order to preserve social tranquility. The point is that here a private act has public consequences. Whether or not the public's intolerance is rational is a secondary question.

Most American cities and states now have laws against the sale of drug paraphernalia — special pipes and cigarette papers for smoking ma-rihuana, mirrors engraved with the word "cocaine," and so on. Ostensibly these laws are meant to make drugs less attractive by getting rid of accessories that have a glamorizing effect. It is as if we tried to prevent alcohol abuse by banning fancy wine glasses, bottled mixers, or 5wizzle sticks with drinking jokes printed on them. This sort of law has practically no effect on drug use. Its real purpose is to eliminate an affront to propriety and decency. Closing down a shop that sells paraphernalia is a way to remove an offensive sight from the streets and put an annoying thought out of people's minds; it is like insisting that pornographic bookstores be inconspicuous. This is a clear case in which a law justified as a way of reducing drug use is actually a way of reducing the anxiety caused by excessive public consciousness of drug use.

Two interesting examples of what seem to be pure legal moralism are criminal laws against desecrating the flag and mutilating a corpse. These acts, Mill would have said, neither harm any assignable individual nor violate a specific duty to the public; nor do they injure the person who performs them; but they are very offensive. In one recent model penal code (Schwartz 1971), flag desecration is a crime only if done in public, but mutilation of a corpse is a crime under any circumstances; public - outrage, in this rare case, is assumed to be carried over to private activity. Today it would be hard to get general agreement on outlawing any private behavior that outraged sensibilities less than mutilating a corpse. Many people may regard flag desecration, even in private, as a symbolic threat to communal order, but that does not justify making it a crime. In this view, it also makes no sense to treat private homosexual behavior as a crime, even if many people think that homosexuality is a danger to the corrununity. Drug use might be treated the same way, as the Alaska Supreme Court suggested when it ruled that smoking marihuana in the home was protected by the right of privacy proclaimed in the state con-stitution, but other courts have not followed it.

One underlying reason for our uneasiness about legal moralism is that we lack any traditional, fully accredited source of authority. We have little sense of social solidarity, and existing customs are under constant attack. At one time the worst crimes included sacrilege, heresy, witch-craft — acts that could bring down the wrath of a deity who would destroy the community. In a more abstract form, the idea of natural law justified tlire punishment of acts that injured an objective public good based on an inherent human nature. Modern liberal societies no longer endorse the idea of a universal moral order backed by either a god's commandments or a natural law prescribing the proper ends of humanity. T. H. Green agreed that it was wrong to punish conduct that violated no one's immediate hefts, just because it was thought likely to weaken belief in some divine authority or bring on the wrath of unseen powers. Mill himself, of course, rejected the theory of natural law as well as all divine authority, since he did not believe in a fixed human nature and fixed goods that everyone had to recognize.

But the presumed outrage of the man on the Clapham omnibus, un-supported by any broad principle, is a poor substitute for the anger of a god or an accepted conception of public virtue based on a view of humanity's natural ends. The U.S. Supreme Court declared contemporary community standards to be the basis for judging whether a legal definition of obscenity is constitutional, in effect endorsing something like Devlin's view for this purpose; but its decision satisfied no one, because people cannot agree even on what these standards are, much less on why they should have the authority of law. Devlin's deference to the alleged indignation of the common man merely exchanges a set of serious ideas that are no longer tenable for a set of unconsidered attitudes that never were tenable. The conviction that some acts are destructive to the common moral order becomes dubious when we no longer believe in an avenging deity who has the power to destroy the community; the conviction that some basic human standard is being violated becomes questionable without the sup-port of assumptions about human nature that would justify it. In these circumstances we have to fear that, as Dworkin suggests, a consensus (not even universal) of prejudices and passions — what Mill calls "the tyranny of prevailing opinion and feeling" (1859, chap. 1) — will fill the gap, disguising itself as a moral imperative.

So, the current Iranian idea of a moral order remains much more comprehensive than any available to Western liberal secular societies. Most of us are inclined to say that in its campaigns against rock music (and homosexuality), the Iranian government is legislating not morality but taste. Where to draw the line is the problem; in liberal societies, more and more things once considered moral issues become matters of taste, or, as Mill would say , tastes and pursuits. We have become reluctant to justify laws by what look to us like confused generalizations about the social fabric or the moral order. We are committed to being at least a little more precise about what we want to protect — public health, public tranquility, aesthetic values, productivity, the welfare of children, and so on. Thus, we free ourselves from the difficulties of legal moralism while retaining its main advantage in avoiding the consequences of Mill's principle: the way it extends the notion of long-term effects or secondary harm. If nothing less than the health of the community or the fate of its children is at stake every time someone performs an apparently fimocuous sex act or casually uses a drug, we need no subtle arguments about whether people should be allowed to injure themselves.

Unfortunately, there remains the problem of adjusting legal means to social ends when the harm contemplated is so vague and all encompassing. In discussing issues such as a decline in productivity or public trafiquility, it is almost impossible to distinguish causes from symptoms. Is the behavior that worries us the source of arty social miseries associated with it, or is it just a by-product of them or an ineffectual attempt to cope with them? We are often uncertain how to justify laws for such purposes; unlike most consumer protection laws, they apparently lack a clear, limited purpose and a well-defined effect.

But drugs are much easier to legislate against than music or sex. The greatest authoritarian philosopher, Plato, thought music and sex required government regulation; the ruling Iranian clerics and other successors in the authoritarian tradition tend to agree. Liberal theory, and now also liberal society, have put this in doubt. But all of us except a few libertarian eccentrics think that drug use requires government restraint. A police campaign against popular music looks absurd to most of us, and a government attack on homosexual behavior looks sinister to some of us, but the point of view from which a war on drugs looks absurd or sinister is hard to achieve today.

The familiar explanation is that drugs are not like sex and music; they are poisons that can produce deadly habits as well as accidents and crimes, 'a threat to life and health. No one could say that objections to drug use are merely a matter of taste. Cause and effect seem clear enough here, and tallc about poisoning or pollution is more than a metaphor; the consumer protection analogy is available to fall back on. But then all the questions that made it necessary to consider the issue of legal moralism arise again. Why are drug laws so severe compared to other consumer safety laws, and why are they so much harder to enforce? Why do we make an exception for alcohol? Why do Iranians (and even some Americans) blame dnrgs and rock music for some of the same evils? Why do well-informed people often see less danger in drugs than ill-informed ones? What sort of consumer protection is this? In fact, drugs are regarded as both a 41angerous commodity and part of a disapproved way of life, so they are subjected to two lcinds of moral and legal censure. But the rules appropriate for these two situations are not the same. The problem is circumvented by using a very broad conception of public health.

Selling or using a drug once may seem harmless, but that is an illusion if jt is the start of an epidemic. Some acts may be a surrender to impulses that have no natural bounds and would spread enormously, overpowering and disintegrating us, if they were not confined by law. Since this is not like the ordinary case in which consumers merely need protection from their own carelessness and the rapacity of others, it is hardly a surprise to find studies showing that people who know something about the actual dangers of illicit drugs are often more rather than less willing to use them. For anyone who regards drug use as a plague, that only makes the threat worse and fierce resistance more necessary. The failure of education and information is evidence that drugs disturb the brain, robbing us of reason, and create addictive habits, robbing us of freedom. It might even be better to keep people in ignorance about them, just as we preserve them from contact with plague germs. Popular indignation can then be regarded as a sound basis for public policy even if it is apparently ignorant, for it acknowledges a basic vulnerability.

In its 1980 campaign platform, the Republican party called the U.S. drug problem a "murderous epidemic." Many social problems are described from time to time as "epidemic," and this is a legitimate use of the word, but it still refers primarily to physical disease. It is significant that drug use evokes the image more often and more powerfully than any other social problem. Traditionally, a plague was the vengeance of the gods on a community whose leaders had transgressed the moral law: the plagues of Egypt in the Old Testatnent or the plague of Thebes in Sophocles' Oedipus. Epidemic physical disease can no longer be regarded as merely a sign of moral disorder, because we know its causes, but drug abuse is another matter. By calling it an epidemic, we suggest a public health campaign that is also a moral crusade.

If drug abuse is a communicable disease, and drugs are a menace like the typhoid bacillus or the smallpox virus, the reasons for intervention become overwhelming. The imagery of disease has tremendous social potency. It eliminates most moral and political doubts, since disease has nothing to do with free action. Preventing an epidemic of typhoid presents no moral problems, so why should the prevention of a drug abuse epidemic? The infectious disease carrier, the Typhoid Mary, has to be quarantined, so why shouldn't heroin addicts be locked up? If the persuasion and imitation by which drug use spreads are regarded as a form of infection, the drug as a disease agent, and the drug user as a carrier, freedom and individual desires obviously deserve no consideration. We grant no one freedom to be infected with typhoid. The thought that everyone might perform homosexual acts or read pornography may be troubling, but on this analogy the prospect that everyone might constantly use drugs for pleasure is worse and also more likely — as though half of the population contracted a horrible disease. Worries like Mill's dissolve, because we no longer have to think of the problem as one created by peoPle doing things to themselves that others consider bad for them; instead we can picture an external agent, the drug, invading the individual and social body and spreading irresistibly. Preventing disease has been regarded as a government responsibility for at least a hundred years; that is the mean-ing of public health medicine. The international system for catrol of pleasure drugs developed along with the first international institutions devoted to preventing the spread of infectious disease, and today the World Health Organization (WHO) serves both functions.

Since so many people are so obviously threatened by any outbreak of contagious drug abuse, Mill's claim of freedom to do what affects only oneself cannot be raised. And in this situation we have to act even on remote and indirect possibilities of harm, as we do when we quarantine and vaccinate. Thus public health, like legal moralism, provides a way of emphasizing external causes (the germ, the chug) and consequences (the transmission of illness to others). And if paternalism is acceptable at all, preventing disease is one of the best possible reasons for it. Rawls's rational person, for instance, would certainly want to be protected against anything likely to cause illness; health is a candidate for the status of sprimary good. Even the idea of natural law is still socially effective in matters of health, because all people have similar ends here merely by virtue of their biological humanity.

Above all, the language of disease tends to foreclose the issue of whether there is any benefit in using drugs for pleasure (after all, some diseases produce an unhealthy euphoria in the early stages). Therefore it eliminates all acknowledgment of any important resemblance between drug use and other tastes, pursuits, and indulgences. We know that it would be absurd to speak of a murderous epidemic of mountain climbing. If we can solemnly refer to a murderous epidemic of drug use, we do not have to consider whether marihuana or LSD, for example, actually causes as many deaths as mountain climbing. It also becomes harder to think of marihuana use as something like an eating habit — an indulgence that may be healthy or unhealthy, good, bad, or morally indifferent, depending on the persons, amounts, and circumstances involved, and that is normally subject only to informal social pressures (or Mill's education and persuasion). A community threatened by infectious disease must simply take the necessary measures to stop it.

The vocabulary of public health medicine also permits a smooth transition from physical health to psychological and moral health and finally to social health: The "murderous epidemic" is crime and illness at once, without careful distinction. In this way, consumer safety becomes mixed with morality, and the two different kinds of justification reinforce each other. 'Mill rejected alcohol and other drug control laws on the ground that they gave the majority a legally enforceable interest in every citizen's moral, intellectual, and physical perfection, to be defined as the majority saw fit. A plausible reply is that curbing mass drunkenness is not a form of officious meddling aimed at remaking some people in the image of others; it is a matter of elementary public health as well as social order. Like the idea of order, the idea of health is ambiguous; it has social and moral as well as biological elements. The World Health Organization once defined it as not just freedom from disease but total well-being, physical, mental, and social. By this definition, anything desirable is healthful and vice versa. In effect, it allows positive liberty to be introduced as an aim of government in the guise of public health, sweeping away most barriers against paternalism. This way of thinking about health recognizes its etymological and actual kinship to wholeness; to be cured is to be reintegrated, restored to oneself, and returned to society. But it also allows public health to develop connotations that belong to political philosophy as much as to physiological medicine. At the turn of the century, the German physician and statesman Rudolf Virchow wrote, "Medicine is a social science, and politics is nothing but medicine writ large." This open-ended conception of health has left its sources in biology far behind. It is characteristic of what is sometimes called the "therapeutic state" — a twentieth-century development that undoubtedly would have disturbed Mill.

Besides, disease is the realm of medical expertise; WHO has even officially identified "drug abuse" with nonmedical use. The British Com-mittee on Obscenity and Film Censorship (Williams 1981) pointed out that banning pornography had become hard to justify, not only because the actual effects of pornography are unclear, but also because we do not know who should judge whether those effects are good, bad, or indifferent; there are no experts on moral health, as there are on medical matters such as drug addiction. The committee's faith in medical expertise on drugs is probably too firm; it remains doubtful to what extent drug use really is a medical rather than a moral or cultural problem. But what matters is the effect of a medical definition; it obviates the problem of guarding the guardians, appointing and justifying an authority. In the twentieth century, medicine has finally become a science, so medical control seems to be based on objective standards supplied by the most reliable intellectual authority we have. Medical regulations are,our most stringent consumer protection laws. A free market in this service is con-sidered an invitation to fraud, so practicing medicine without a license is forbidden. The ordinary citizen has no more claim to judge the proper uses of drugs than the theories of biology. Mill did not approve of the taste for opium, but he objected to laws making it impossible to indulge the taste without a doctor's prescription. Today, taking opiates is no more considered a matter of taste than using an x-ray machine. To do it without supervision is simply a dangerous misuse of an instrument valuable only for certain purposes understood by the medical profession. And the rules against talcing risks in medicine are much stricter than the rules against taking risks in the pursuit of pleasure or ambition. If the sport of boxing came under medical control, it would be banned immediately.

H. L. A. Hart criticizes Mill for assuming implicitly that most people are "rational middle-aged men with settled preferences" (Hart 1963, p. 33) — in other words, that they resemble Mill himself. Hart thinks that Mill's antipaternalistic rule would enhance the freedom and welfare only of a minority, perhaps an elite. Medical control is an interesting variation on this idea. It would be convenient if we could find a way to refuse drugs directly to only those people who would use them impulsively or unwisely, and make them available to others, but that is administratively impossible. Instead, control of drugs is turned over to an elite of presumably rational and mostly middle-aged men and women who are assumed to know how other people should use them. But their authority is not derived from being generally sensible and well-balanced; it comes first from professional knowledge and only secondarily from personal character and wisdom.

The history of ordinary medical drug regulation shows that even this authority becomes more and more restricted as the presumed objectivity and certainty of medical knowledge increase. The first federal drug law, the Pure Food and Drug Act of 1906, was hardly a consumer protection law by modern standards of drug regulation. Its main purpose was to guarantee fair value by prohibiting "false and misleading" labels. Mill would not have disapproved, since he agreed that people need protection against fraud. In a 1910 Supreme Court case, Justice Oliver Wendell Holmes pronounced that the phrase "false and misleading" could apply only to claims about the contents of the package, not to claims about its usefulness. On that issue, he said, opinions are far apart and consumers should be allowed to judge for themselves.

lu 1910 there were very few effective drugs and very little solid evidence on which ones were useful or harmful in which situations. Today dis-agreements about these matters are less common, and there are scientific procedures — especially double-blind clinical trials — designed to resolve any differences. Lay persons now have little more to say about whether most drugs work than about whether opium use is a matter of taste. The Food, Drug and Cosmetics Act of 1938 was the first federal law requiring sellers of medicines to provide directions for use and proof of safety. These provisions were soon interpreted by administrative regulation to require doctors' prescriptions for many drugs; the assumption was that laymen could never properly follow directions for using those drugs on their own. In effect, the law established a distinction between prescription and nonprescription drugs for the first time. More elaborate restrictions introduced in the 1950s and 1960s gradually took decisions on which drugs could be prescribed away from doctors and transferred them in-creasingly to the government. Not even ordinary doctors, much less lay persons, would be allowed to choose, even when they were fully informed. The whole development could be described from one point of view as a vast extension and transformation of the principle allowing laws to prevent fraud — an extension to the point where any doubts about paternalism are dissolved. It would have been impossible without the conviction that appropriate experts can determine with scientific confidence which drugs work and which do not.

Medical terminology and medical metaphors seem natural in dealing with psychoactive drugs. Most ordinary therapeutic dnigs are dangerous enough to cause illness as well as cure it. (Drugs taken for pleasure are sometimes said to produce a special kind of illness — addiction or de-pendence.) Most pleasure drugs also have regulated therapeutic uses. So it is plausible to apply the same strict rules. We do not have to worry about whether marihuana smoking is more dangerous than boxing or flying a small plane; medicine is not in the same conceptual realm as voluntary tastes and pursuits.

But medicalization can never be the whole answer for pleasure drugs. The regulation of these drugs has a separate historical origin and remains distinct today. Prohibition of so-called narcotics began long before the government required prescriptions for ordinary therapeutic drugs. Consumers have not had the right to choose opiates or cocaine since 1914 or marihuana since 1937. The government was telling doctors that using opiates to maintain an addict was not a legitimate medical practice long before it started to substitute its judgment for the doctor's about any other kind of drug therapy. And the usual forms of medical control have, of course, always been inadequate for the drugs now designated as controlled sub-stances. Even when they have therapeutic uses, prescriptions arérecorded and limited by law, and doctors may be called on to justify their use to the federal government.

The language of medicine was introduced in the first place to malce the special severity of drug controls plausible where justifications based on public morality or even simple consumer protection might ha'ite been inadequate. But since the usual forms of medical control are obviously too mild, we continually turn back to the other justifications to supplement the medical one. The circle is never made explicit, since the transitions are blurred by the ambiguity of a term like "public health"; this special public health problem includes psychological, social, and moral health. So the use of drugs for pleasure is believed to present three kinds of threat to human welfare. It is an offense to morality, or at least a danger to public tranquility, productivity, and so on; in some ways, it is also like an epidemic disease; and in some ways, it resembles the ignorant use of a dangerous instrument like a chainsaw. Social attitudes and legal regulations conform to each of these three analogies in different ways, and each one reinforces the others at weak points to supply reasons for stricter controls. For example, viewed as a way to enforce a consumer protection measure (or promote public health), spraying marihuana with the dangerous weed killer Paraquat would be absurd. But viewed as a severe deterrent to inunoral behavior, it makes excellent sense. We officially reject the idea that drug use might be a risky taste or pursuit that nevertheless sometimes has value for some people, like climbing mountains or eating rich food. Marihuana smoking can be seriously described as an epidemic, but to call mountain climbing "getting high" is only a joke.

This way of using analogies is possible only because drugs are assumed to do little or no good. Since the pleasure they provide is seen as trivial (if not bad in itself ), freedom to use them is trivial, too. Where drugs are concerned, we all have to be treated like children, unable to judge our interests well and in need of guidance from coercive rules. Making our own choices about drugs is not thought to develop richness, diversity, or unity in human experience — on the contrary. A utopia in which no one wants to do anything so dangerous as climbing mountains is absurd, but a society in which no one wants to use drugs for pleasure is still regarded as a plausible ideal — the drug-free society.

One way to avoid this conclusion is to redefine the value of drug use so that it looks almost as good as mountain climbing. In the late 1960s and early 1970s, there was much talk of consciousness expansion. Today we sometimes hear about a natural desire to alter consciousness that appears even in children and that could be said to justify the use of drugs (Weil 1972), just as a natural desire for adventtur or competitive achievement justifies mountain climbing. There would be no need to seek pathological causes. In practice, altering consciousness to diversify experience (in the abrupt way that drugs do it) is not generally recognized as a natural desire with value of its own. But in one interesting exceptional case, U.S. law has recognized the use of a drug for other than medical reasons as an activity with intrinsic value.

Members of the Native American Church, an Indian group, are allowed to take peyote in their religious rituals. Here federal courts have found a fundamental right of the individual that overrides a state interest in suppressing nonmedical drug use: the guarantee of religious freedom in the First Amendment to the Constitution. In other words, drug use has to be more than a pleasurable taste or pursuit before the law will allow it. To refute the presumption that nonmedical drug use is negligent, ignorant, and generally worthless, there must be overwhelming evidence that the drug users know what they are doing, consider it important in their lives, and believe seriously in its intrinsic value. But even that is not enough. The courts have made it clear that they will not accept merely individual religious beliefs (much less consciousness expansion) as a justification for drug use, and they have said that they will scrutinize very skeptically the claims of any new organized churches. The drug must be not only religiously important to its user but also an essential part of a traditional rite with a communal significance. So far, the exception made for the Native American Church is unique. It is as though mountain climbing were regarded as generally so dangerous and useless that climbers would be fined and jailed unless they could prove they were making a pilgrimage to a holy site on the peak certified by an established church.

The peyote eaters have been allowed to escape from the standard system of categories. Their drug use is not for pleasure, it is not exactly medicine, and it is not illness or impulse or error. The religious context makes this an obvious case for Mill's view that people should be free to shape their lives, guided by what is significant for them. But there is also a special social context. Peyote eaters are thought to deserve respect because they demand something different from the abstract liberal individualist freedom to do what they want if it injures only themselves. The idea of liberty involved is inseparable from the maintenance of a community and a common way of life. We are implicitly endorsing a kind of positive liberty in the guise of religious freedom.

The Native American Church is one small exception. Alcohol is a different kind of exception. We know that it is not absurd to treat drug use as a matter of taste, because that is the way we have always regarded alcohol. It is not just that alcohol is legal and most other pleasure drugs are illegal; our whole public discourse about alcohol, even that of pro-hibitionists, is different from the way we talk and thinlc about other drugs. No one pretends that alcohol prohibition can be treated as a consumer safety measure. For example, during the campaign for Prohibition, it was understood that a ban on alcohol would require a constitutional amendment, unlike other drug laws and consumer laws in general. To begin outlawing opiates and cocaine, all we needed was a law based on the taxing power granted to Congress in the Constitution (the Harrison Act of 1914 was formulated as a revenue measure, because most police powers were thought to be reserved to the states by the Tenth Amendment). But possession of alcohol was never a crime even under the Volstead Act, and since repeal we have relied almost entirely on education and appeals to self-restraint instead of paternalistic coercion in dealing with alcohol problems — a limitation unimaginable in the case of any other drug. Alcohol has never been under medical control either, except briefly during Prohibition. We do use disease analogies in talking about alcohol abuse, but the emphasis is on the susceptible individual and not the irresistible infectious agent. Few people, except those in Moslem countries, dream of an alcohol-free utopia, and even fewer want to impose it by law. Alcohol escapes the net of analogies used to control other drugs.

The most common argument used to justify the inconsistency is that cohol prohibition, unlike other drug prohibitions, just does not work. t is true that even in a completely free market, alcohol might be the most popular drug; that makes banning it harder. But prohibition may well 'reduce the consumption of alcohol and the harm done by drinking; we do know that American alcohol use declined in the 1920s, and the incidence of diseases such as cirrhosis of the liver dropped sharply (Burnham 1968— , 9). Prohibition is also said to cause more harm than it prevents, but if that is true, it is equally true of other drug laws. Alcohol prohibition in the 1920s and marihuana prohibition today may have produced the same nasty side effects: costs of arrest and punishment, growing disrespect for the law, organized criminal violence, police corruption and oppression, poisonous adulteration, and misrepresentation. The parallels are ridiculously precise, including the tendency (in the past, if not more recently) for research to be monopolized by prohibitionists and the appointment of national commissions (the Wickersham and Shafer commissions) that investigated the problem and came to self-contradictory but cautious and 'therefore politically acceptable conclusions. In fact, allowing for differences in the size of the original problem, alcohol prohibition probably worked just about as well (or badly) as present drug prohibition laws work. Repeal came not because prohibition was totally ineffective, but because we decided — although we seldom express it this way — that we wanted the igeasure of convenient, legal alcohol more than we feared an increase in dnmkenness and alcoholism. It is still unthinlcable to make the same kind of balancing judgment about any other drug, even to come to a different conclusion.

We concede that alcohol is a very dangerous substance and creates a vast health problem, while recognizing that it can also be a harmless indulgence. But apparently the strain of tolerating the ambiguity is too great, because we are unable to do the same for any other drug, even when there is little evidence that the drug could ever be as dangerous as alcohol. In public discussions, drugs are a dangerous commodity, a public health problem, and a moral menace that requires the full force of the law. That is the official view. But unofficially it is different; in polls a large part of the population says that drug use should be a matter of individual choice (Glaser and Snow 1969). What may seem to be hypocrisy here is largely confusion and doubt. Drug use is a difficult social policy issue because it is so complicated and ambiguous. It is a hard test for modern ideas about freedom and modern ways of differentiating classes of behavior. Many analogies seem applicable to it, and we easily fall into inconsistencies when thinking about it. Before considering historical and other explanations, it is important to exatnine one more topic that is closely connected with the theoretical issues discussed here: addiction or dependence.