The most important justification for strict legal and social controls on drugs is dependency or addiction. This is the kind of drug use that produces the most serious effects on health, productivity, and family life. Even more important, it provides the best reason for saying that the drug user is not free, and that anyone exposed to the drug may lose personal freedom. Respecting a person's freedom may not require respecting his or her desires if those desires are addictive or may produce a dependency. Besides, addiction and dependence are what make us think of drug abusers aa sick and drug use as primarily a medical problem rather than a mere taste or pursuit. Dependency makes the drug user resemble a child or a patient, who can justifiably be deprived of autonomy.

But human lives are inconceivable without habitual actions; virtues and vices are habits; our personalities, and the very continuity of our , selves, are partly constituted from habits, in the sense of learned dispositions to certain ways of responding and behaving. There is no sharp line . between acts determined by choice and those determined by habit. And any self-destructive or immoral habit can be pictured as a kind of en-: slavement, in accordance with the view that only positive liberty is true liberty:

. . each of these petty beings [habits] held secretly a Chain in her Hand . . . though they were always willing to join with Appetite, yet when Education kept them apart from her, they would very punctually obey Command. . . .

Though they grew slowly in the Road of Education, it might however be perceived that they grew; but if they once deviated at the Call of Appetite, their Stature soon became gigantic, and their Strength was such, that Education pointed out to her Tribe many that were led in Chains by them, whom she could never more rescue from their Slavery . . . but with little effect, for all her Pupils appeared confident of their own Superiority to the ' strongest Habit, and some seemed in secret to regret that they were hindered from `, following the Triumph of Appetite. . . .

Others were enticed by Intemperance. . . . I observed that the Habits which hovered about them soon grew to an enormous Size. . . . Habit had so absolute a Power, that even Conscience, if Religion had employed her in their Favour, would not have been able to force an Entrance. (Johnson 1950, pp. 150, 157)

"Intemperance," in this eighteenth-century text, means what we call "alcoholism." The term "intemperance" was used mainly to refer to habitual drunkenness even before the term "addiction" came to mean chiefly opiate abuse. Obviously people have always been conscious of the unusual strength and devastating effect of some drug habits. But it is only recently that we have come to think of those habits as somehow different in principle from others.

It is not easy even to decide whether the issue is an empirical one or merely an argument about words. The philosopher W. B. Gallie has introduced the idea of "essentially contested concepts." The criteria for applying these concepts are multiple and involve moral and social eval-uations, and the relative importance of the various criteria is recognized by users of the concept to be unsettled. The life of the concept is debate. Like such political ideas as democracy and tyranny, addiction or dependence qualifies as "essentially contested" in this sense. That values are involved is obvious; we do not usually call any habit an addiction or dependency (except as a joke) unless we mean to say that it is harmful. It is -probably not so clear that the criteria for application are unsettled, since the prevalence of medical terminology in the field may make it appear, falsely, that drug dependence is as clearly defined as cancer, smallpox, or even schizophrenia. But in fact, today as in the past, even medical authorities have not been able to agree on what drug dependence is. In one recent collection of essays on the subject, twelve psychological theories, eight biological theories, twelve sociological theories, and fifteen mixed theories are offered, all with at least some claim to respectability (Lettieri et al. 1980).

An addiction once meant almost any strong habitual inclination, as in this line from Shakespeare's Othello: "Each man to what sport and revel his addiction leads him." It was common to talk of an addiction to gardening or theater going without deliberate humor or any sense of incongruity. But to many people today, addiction means simply a mysterious and utterly debasing enslavement to certain dread chemicals. Other meanings are felt to be secondary, metaphorical or jocular. This change has come about partly because the concept of addiction has been medicalized (with few corresponding advances in the discovery of either causes or solutions) And yet, ironically, the medical profession is now beginning to abandon the term; it is no longer applied, as it once was, to almost any habitual drug use that incurs social disapproval.

One medical definition of addiction that still has some currency is the following: a condition induced in certain higher manunals by chronic administration of central nervous system depressants like alcohol, bar-biturates, and opiates, in which a gradual adaptation of the nervous system to the drug causes a latent hyperexcitability that becomes manifest when the drug is withdrawn and produces physiological symptoms that are interpreted as a physical need for the drug. This definition implies no moral or political attitude, and it does not intimate anything implausibly horrible and debasing (the much-publicized heroin withdrawal reaction at its most intense is no worse than a bad case of flu, and exaggerating it only serves the interest of the addict's self-dramatization). Emphasis on the withdrawal reaction, which is physically identifiable, at least keeps medicine out of the business of judging which drugs are detrimental to society or trying to solve the problem of free will by distinguishing between compulsion and desire. But there are some problems even with this definition. Other drugs — stimulants, caffeine, and nicotine — produce various physical withdrawal reactions in many people, and withdrawal from a person — the breakup of a love affair, for example — sometimes produces similar symptoms. But the main problem is that physical with-drawal reactions are simply not one of the most important causes or (with the exception of some high-level barbiturate addictions) one of the most serious consequences of drug abuse. If they were, it would be more plausible to define it as a purely medical problem.

Pharmacological tolerance is another symptom that often accompanies addiction. It is usually described as an adaptation of the nervous system to the effects of a given amount of a drug that makes it necessary to keep taking more to get the same effect. Tolerance is most conspicuous in amphetamine abusers, alcoholics, and heroin addicts, but most drugs will produce it if they are taken often enough for a long enough time. Again there is a question of whether this reaction is something specifically induced by drugs; all routine pleasures tend to pall and may have to be revived by an injection of more of the same. The call for madder music and stronger wine does not distinguish between the wine and the music. Tolerance to drugs also varies a great deal with the individual and the particular effect; for example, some amphetamine abusers take twenty times the amount an ordinary person could tolerate, but the drug can be given for years to a person suffering from narcolepsy without any need to raise the dose.

Little is known about what, if anything, distinguishes drug withdrawal reactions and drug tolerance from other forms of compensatory response and adaptation. One theory is that withdrawal reactions occur because a drug has taken over the work of a chemical normally made by the body itself and has caused the mechanism that produces it to atrophy, like an unused muscle. Another theory is that in supplying chemical stimulation to nerve cells, drugs produce physical changes that increase the nerve cells' demand for stimulation, creating a need that normal body processes can no longer serve. But analogous theories might be needed to account for the symptoms of habituation to things other than drugs; body chemistry has to change in either case. Nor do we have evidence that the induction of drug tolerance and withdrawal reactions is in general so much simpler, more reliable, and more mechanical than other forms of habituation — so much more independent of what we usually call free will and choice _— that it deserves to be treated quite differently for moral and political purposes.

As the term "addiction" has come to be used in medicine less often and with a more restricted meaning, "dependence" has taken its place in many official formulations. In 1969 WHO's Expert Committee on Drug Dependence defined it as follows: "A state, psychic and sometimes also physical, resulting from the interaction between a living organism and a drug, characterized by behavioral and other responses that always include a compulsion to take the drug on a continuous or periodic basis in order to experience its psychic effects, and sometimes to avoid the discomfort of its absence. Tolerance may or may not be present" (World Health Organization 1969, p. 61). 'The apparent neutrality of "dependence" seems attractive as a way of avoiding premature theoretical commitments, and the word is used by so many people for so many purposes that it is unlikely to suffer the fate of that captive of the drug-abuse vocabulary, "addiction." But many are still dissatisfied. Some authorities, who might be called humanists, think that it suggests a kind of enslavemont and falsely implies that people who use drugs are not choosing to do so. They point out that in the vague WHO sense, drug dependence is no different from dependence on television, a religious ritual, a parent, or even trousers. A life centered on drug use may be unreasonable and self-destructive, but so are many other devotions and commitments. Explanatory terms like "compulsion," "craving," and "overpowering need" apply just as well to love of chocolate cake or, for that matter, to love of another human being.

Some writers who take this attitude consider themselves defenders of pleasure against a life-denying puritan morality; others see themselves as defenders of individual liberty against state power and medical technocracy. Thomas Szasz, for example, insists that a person is neither more nor less responsible for misuse of psychoactive drugs than for any other bad habit or vice. For him the idea that some chemicals have a mysterious power controllable only by coercive authority removes the main actor from "the drama of temptation and restraint" and makes it a meaningless exercise (Szasz 1974).

The trouble with "compulsion," "craving," and other such expressions is that they suggest a simple state of mind that produces a simple, uniform pattern of behavior. But the behavior of drug abusers follows no such pattern; it is impossible to define when an established routine becomes a compulsion or a strong desire a craving. Alcoholics, for example, when asked why they relapse, usually blame it not on craving but on anger, frustration, and social pressures. "Loss of control" has become less popular as a way to describe what happens when alcoholics drink, because some of them can take one drink and stop if the time, place, and social setting are right (Paredes et al. 1973), and to say that alcoholics lose control only when they drink abusively would be tautological. Heroin addicts tend to lose their craving when they are in prison or some other place where they cannot expect to find the drug; they often feel it again when they return to their old environment (Meyer and Mirin 1981). Not only do even the most severe addicts and alcoholics exercise some control but, contrary to myth, the great majority of opiate users are not addicts at all (Jacobson and Zinberg 1975).

Alfred Lindesmith developed a theory of opiate addiction that defined craving as the feeling that arises when an addict interprets abstinence symptOms (withdrawal reaction) and attributes them to the drug (Lindesmith 1947). But it turns out that addicts tend to take more than they need to relieve the physical symptoms of abstinence, and most of them come to desire opiates with an intensity that can be called a craving before they are physically addicted; in fact, it is this desire that causes addiction, rather than the reverse. Craving for stimulants even more obviously develops without physical addiction. Craving can be defined in a commonsense way as a desire that is abnormally strong or hard to change, or that produces an abnormally intense reaction when it is not fulfilled. But drugs are not the only object of that kind of desire. The authorities we have called humanists therefore tend to regard it as obvious but irrelevant that people are likely to want to go on doing something that gives intense pleasure. As Szasz puts it, to talk about a euphoriant without the potential to produce craving, dependence, or addiction in some sense is as absurd as talking about a flammable substance that does not ignite. The humanists seek other reasons for the persistence or intensity of the desire, and for moral approval or condemnation of it, in personal history, environment, and culture.

On the other side are researchers who are trying to develop a scientific model of drug dependence; often they are behavioristic psychologists who want to measure the dependence in experimental animals with the help of their favorite conceptual apparatus of reinforcement schedules and operant conditioning. From the point of view of these scientists, the trouble with WHO's language is not its social implications, or some insult to human dignity, but its vagueness, which permits too many debaters' points in opposition. They would like to clean up and shave down this shaggy definition for laboratory purposes. For them, a drug produces dependence if it is taken repeatedly by animals in cages. Differences between animals and people, between one animal and another, or between one person and another are secondary. References to desires, cravings, compulsions, and other "mentalistic" phenomena are not allowed. Moral judgment, implicit or explicit, is forbidden as unscientific. Here we are beyond freedom and dignity, and everything must be cashed in quantitative, observational currency.

Since behaviorism has no use for the idea of free action, it tends in its own way to deny any special features to drug dependence. Drug habits are said to be produced, like other habits, by repeated stimulus and response. Objects differ only in their reinforcing power for a given organism in a given situation. Here is a typical behavioristic definition of drug dependence by M. H. Seevers: "Repeated use of psychoactive drugs leading to a conditioned pattern of drug-seeking behavior. A characteristic predictable and reproducible syndrome is associated with each drug" (Seevers 1972, p. 17). Experiments with laboratory animals do reveal distinctive patterns. Rats and monkeys will not repeatedly inject the psy-chedelic drug mescaline or the antipsychotic drug chlorpromazine, and these drugs do not produce what is called drug dependence in human beings. They will inject morphine, cocaine, amphetamines, barbiturates, alcohol, and caffeine. The pattern for opiates like morphine and codeine is a gradual increase to a dose that is then held constant, while the interval between injections varies. When the injection machinery is disconnected from the drug supply, the animal keeps trying to get more at a low rate for a long time — up to months. The pattern for cocaine and amphetamines is a cycle in which rapid injection of large amounts is followed by exhausted abstinence. When the machinery is disconnected, the animal keeps pressing for the stimulant at a very high rate, but only for a few hours before it gives up (Thompson and Pickens 1970). Significantly, human opiate and stimulant abusers may behave more or less the same way in analogous situations (Griffiths et al. 1980).

There are authorities who regard behavioristic explanations as not scientific enough; they object to definitions "based on social and behavioral criteria, not on any characteristic biologic, biochemical, or neurophys-iological aspects of [drug] use" (Fink 1972, p. 384). They are impatiently awaiting a concept of drug dependence soundly based on physical tissue responses. So far even physical withdrawal reactions have not been fully explained in terms of tissue response. But psychoactive drugs do closely resemble the chemical neurotransmitters that are manufactured by the body to pass messages between nerve cells, and this suggests another explanation of drug dependence. A psychoactive or euphoriant drug is what biologists call a "supernormal stimulus." A herringcatcher will hatch a gull's egg in preference to its own, because it has an innate brooding mechanism that draws it toward the biggest egg in sight. Drugs apparently stimulate brain cells as gull eggs stimulate the herringcatcher, with a similar potentially maladaptive effect. They could be regarded as a neu-rophysiological shortcut that evades the environmental adaptations usually needed to obtain pleasure. Drug dependence has been called an "artifical drive," with the same kind of power as primary biological drives such as hunger and sex — not necessarily uncontrollable or unmodifiable, but something more than an ordinary habit (Bejerot 1972).

To the humanists, "conditioned pattern," "predictable syndrome," and "artificial drive" have the look of fighting words, weapons in a war between the two cultures. The main procedural objection is that the conditions of behaviorist experiments are too different from any encountered by animals in the wild or people in society. The animals are caged and under restraint; they have few sources of satisfaction except drugs. The conditioned pattern may be a laboratory artifact, and the animals might act quite differently in a situation that offered a variety of dangers and opportunities. Even in laboratories, the reinforcing properties of drugs vary from one animal to another. And most free-running animals in a more or less natural environment apparently do not use drugs in an  intensive and compulsive way even when they are made available without limit (Alexander et al. 1978).

One- of the most interesting animal experiments offers a particularly strong challenge to the idea of a simple, uniform pattern of stimulus andresponse attached to certain reinforcing chemical substances. Rats and other animals were kept at 80 percent of normal body weight and forced to press a lever for food, which was not delivered immediately but after a fixed or variable period of time. When the interval between deliveries of food pellets was very short or very long, they drank water normally. But when the interval was intermediate, they dranlc vast quantities of water in a compulsive, ritualized way. The thinner the animals were, the more of this compulsive behavior they exhibited. Putting them on a liquid diet or delivering water directly to their stomachs through a tube made no difference. After a while, the animals would even work to get the excess water. The result was similar when cocaine, avoidance of an electric shock, or water was substituted for food as the original intermittent reinforcer. The type of compulsive behavior varied, depending on the opportunity. A pigeon would attack another pigeon under restraint. When alcohol was available, rats would become chronically intoxicated, although they normally show only a mild interest in alcohol.

The experimenters interpret this excessive behavior as comparable to displacement activity. Limited quantities of an important commodity are being delivered at rather long intervals; the animal is tempted to escape from the situation prematurely, and delays action by a compulsive ritual. Since everyone is subject to frustration by nature and society, we are all on "a set of complex intermittent schedules" of reinforcement, which may produce "adjunctive behavior" unconnected with the immediate source of the frustration. Opportunities in the environment determine whether this behavior will take the form of irrational violence, drug abuse, or other excesses. The pattern is associated with a situation rather than a substance (Falk 1981).

In any of these cases, distinguishing between the laws of nature and the laws of culture is not easy. All drugs act physically on the brain, but the brain's perceptions of reward and punishment, pleasure ar021 pain, depend largely on what we think, and that depends largely on our culture and the company we keep. At some point, the brain structure common to all mammals, or even all human beings, has to give way to personal histories and social conditions. Even laboratory animals, so long as they are pressing levers to inject drugs in an artificial environment fkee of competing interests, dangers, and exigencies, are in a sense culturally determined to crave drugs or to be psychologically disturbed. The limitations of ideas about psychoactive drugs based on animal experiments are amus-ingly illustrated by the fact that these animals show no interest in tetra-hydrocannabinol, the active ingredient of marihuana, although international treaties classify it as a dangerous "narcotic" with dependence-producing powers that have to be coercively controlled. The neurophysiological effect of a drug's chemical action is usually necessary to produce the habitual inclination to use it, but this is not enough to sustain a coherent concept of drug dependence. In the real world, drug abuse has everything to do with psychological and social problems.

Looking at individual psychological problems, we come upon the notion of the addictive or dependence-prone personality. This label has been used, confusingly, both to complement and to contradict the notion of a dependence-producing drug. It is said that only an individual with an inadequate personality is inclined to take a bad drug (especially opiates) or that a deficient person will make bad use of any drug. There may be personalities, as well as drugs, associated with drug dependence; some people, for a variety of reasons, seem to need a drug's influence to gain respite from their troubles or surcease from pain. But it is not true that only and all people who have some special characteristics identifiable in advance will use or misuse drugs. For example, in studies of alcoholics, it appears that some are antisocial personalities (Robins et al. 1962); some tend to be depressed (PaIola et al. 1962); some are concerned with power (Williams 1976) and others with dependency (Barry 1976). As for heroin addicts, no single variable of personality or circumstances predicts who will become addicted (Kandel 1978).

In other words, dependence proneness is not a useful diagnostic category (accordingly, the American Psychiatric Association no longer classifies drug dependence as a personality disorder). We cannot tell who will take to the habitual use of which drug. It is too hard to arrange control groups or compensate for the effects of differential availability. In any case, to impute a psychological deficiency to the drug user is circular reasoning if the only symptom of this tendency to drug dependence is precisely the habitual use of an illicit drug. Szasz points out that it is absurd for Ernest . Jones to regard Freud's giving up cocaine as evidence that he was not an addictive personality, while ignoring his far more serious dependence on nicotine in the form of cigars. And William Burroughs commented:

In Persia where opium is sold without control in opium shops 70% of the population is ' addicted [this is a vast exaggeration]. So should we psychoanalyze several tnillion Persians to find out what deep conflicts and anxieties have driven them to the use of opium? I think not. (Burroughs, 1956, p. 127)

Social policy can validate judgments about the personal weakness of drug users. Given existing laws and attitudes, it might be true that people . who want to use heroin in the United States ought to have their heads examined, but the question is whether they should be called weak rather than daring to the point of foolhardiness. The serious argument here is that whatever the weaknesses and inadequacies of drug users, the idea of a private transaction between an aberrant personality and an overwhelm-, ing chemical compound is a myth. If, like most heroin addicts, a drug user is not permitted to lead a normal life, his or her inadequacies will appear magnified. The social conditions in which a drug is available determine the kind of person who will use it and the view that will be talcen of his or her personality. As one sociologist has pointed out, if many opiate addicts were middle class in the 1890s and very few were in the 1920s, it was not the personalities of middle-class people that had changed but the social meaning and function of addiction (Duster 1970, p. 156).

There is another way in which the diagnosis of drug use as a symptom of personal inadequacy can be misleading. Learning to use a socially acceptable drug like alcohol is part of growing up; it is one of the symbols and rituals of adulthood. Marihuana, on the other hand, was until recently identified with youthful rebellion, that is, with immaturity. The rituals of marihuana use therefore had a different value, and one of the worries of parents whose children used the drug was that they were not growing up in the socially approved way. From this uneasiness, it was a short step to the judgment that anyone who used this "wrong" drug, following a slightly different ritual and passing into a slightly different social world, was emotionally immature, weak, unable to face problems — even "fleeing reality." Users of an unfamiliar drug are likely to be seen as menacing if they are far enough away, inadequate if they are too close to home. So a psychological judgment is made on a social situation.

The word "compulsion" in the WHO definition of drug dependence suggests both a pharmacological mechanism and a drive coming from somewhere in what Freud called our "inner alien territory." But in practice these cannot be separated from social situations. Carl N. Edwards describes a patient who became "addicted" to a placebo pill; he demanded dose increases and developed craving and abstinence symptoms (Edwards 1974). This extreme case illustrates in parody the importance of the social context, because here only the institutionalized relationship between doctor and patient produced and sustained the dependence. Even in the more ordinary case, social rules and customs may define the consequences of using a drug; the old word "addiction" too often meant little more than "the consequences of opiate use," many of them produced by laws making it necessary to conunit crimes to pay for the drugs. If someone deeded a barbiturate or whiskey to fall asleep every night, this habit was not called addiction mainly because it was not as expensive as injecting weak solutions of heroin (many so-called addicts do not use enough of the drug to produce a physical withdrawal reaction) or as likely to lead to prison, and therefore did not disrupt the drug user's life or society's routines so much. If medical terminology is used mainly to impose a moral judgment on a social situation (especially one created by the law), it is in danger of becoming meaningless.

It is hard to overstate how much social conditions influence the symptoms of drug dependence. In one study a group of alcoholics was placed in a hospital ward and each alcoholic, one at a time, was allowed to drink as much as he wanted for two days once every two weeks. The drinking was supposed to be done alone and in an assigned area. The result was that the alcoholics drank only where they were allowed to drink and stopped when told to stop without complaint. They even said that they did not enjoy the drinking very much. The supposed chemical tyranny of alcohol was easily suspended when it became dissociated from familiar bars and drinlcing buddies (Paredes et al. 1973).

Thomas Szasz is the most extreme advocate of the view that drug addiction or dependence is a social rather than a physical or psychological fact. He writes: "Addictive drugs stand in the same sort of relation to ordinary or nonaddictive drugs as holy water stands in relation to ordinary or nonholy water" (Szasz 1974, p. xvii). An addictive drug, in other words, is not a particular kind of chemical but something that has been labeled as such by an accepted ritual. Szasz even argues that addiction to a drug is in principle no different from a foreigner's "addiction" to speaking English with an accent, another habit that is hard to break. And he refers to abstention from drugs (or dieting instead of overeating) as another form of addiction. We should not allow these polemical exag-gerations to put us off, even if Szasz does mean them literally. It is not necessary to adopt his radical individualism and the associated classical liberal view of the functions of the state and the medical profession (he would impose no restrictions at all on drug use and sale) to recognize that concepts like drug dependence may encourage people to avoid taking responsibility for the consequences of their habits by treating them as %external forces rather than as part of the self. It also makes us more inclined to use an external countervailing force, and to apply it to the drug itself and the drug users rather than to other sources of misery and mischief in the environment. We avoid difficult decisions about whether the use of a particular drug in a particular way is a good or bad habit, and we do not permit the users to decide either, since, being dependent or addicted, they are by definition unable to make reasonable choices.

Drug dependence, then, involves pharmacology, individual psychology, and social background in varying proportions. This is more than a com-monplace, because it shows that emphasis on the drug itself is often wrong. In its latest diagnostic manual, the American Psychiatric Association reserves the term "drug dependence" for what we have called addiction: tolerance or a physical withdrawal reaction. Other problems involving drugs are classified as organic mental disorders (acute drug intoxication, drug-induced psychosis, etc.) or substance use disorders (most habitual misuse of drugs). Thus psychiatrists now avoid the use of "dependence" in a way that implicitly judges any habitual or persistent drug use as abusive or excessive by its very nature.

But even if drug dependence is not a very coherent concept for analytical purposes, anyone who has tried to give up smoking cigarettes will testify that something is going on that requires that description. Attachments to drugs are too often felt to be unusually intense and yet somehow separate from the center where plans are made — the executive ego. Consider the following example of incipient alcoholism: A retired woman, aged sixty-four, living by herself, has begun to take sherry whcn alone and is now drinking two bottles a day. She has made a rule that she will not drink before 10 AM, but she feels shaky in the morning and sometimes finds it "difficult to hold out." She comments, "It's all so silly, I can't believe it's me" (Royal College 1979, p. 41). And drugs are often personified, as though they were agents with a will of their own: Amphetamines are "the Man" to a truck driver or football player who uses them at work; cocaine is "Lady" to someone having a romance with it; alcoholic drinks are "John Barleycorn" or "Demon Rum." The drug habit is felt to be something outside the self, or a fragment of the self that has become detached but is still on peculiarly intimate terms with it.

However the emphasis is distributed among biological, personal, and social causes, the common theme is imbalance, fragmentation, loss of wholeness, lack of internal direction. Things fall apart, the center cannot do its job of regulation. A part of the person somehow takes control, and the habit somehow both is and is not oneself in action: the piradox of habit. Members of Alcoholics Anonymous are taught that alcohol is an external power enslaving them, yet at the same time that being an alcoholic is their identity. Alcoholics who experience a conversion and give up drinking are not just dropping a bad habit; they are undergoing self-surrender, self-renewal, self-transformation. The American Psythiatric Association, as we saw, has narrowed the application of "dependency," but its broader definition of substance abuse still includes "inability to reduce or stop use," which implies feeling one's own impulse as an alien power.

The main effect of compulsive drug abuse is to make behavior simpler and more rigid; the adaptive repertoire is narrowed, and experience becomes less rich and diverse.* (The extreme case is that of a caged laboratory animal injecting drugs.) But for some people, this is a kind of solution. A heroin addict may be trying to reduce all the problems of life to a single one, so that it will no longer be necessary to make fundamental choices or be subject to ordinary emotional vicissitudes. A life is given structure by repetitive acts that produce an artificial stability. Addicts can be devoted as much to hustling — the daily routine of getting the next fix — as they are to the drug itself. It becomes a way of life. So the chemical action of the drug is real, but its overwhelming power is a projection of a quality that belongs to the drug abuser. The same loss of adaptability ican be found in any monomania — pathological gambling, "mainlining Jesus" in a religious cult, even all-consuming devotion to a beloved person or a gaine like chess. Avoiding addiction or dependence means maintaining " balance by caring about more than one thing (or person). So addiction or dependence is a matter of degree; the ideal of autonomy can only be approached, never absolutely realized (see Peele and Brodsky 1975).

This is another way of stating the political dilemma that both positive and negative liberty seem to be fictions. If reasonableness is produced by a certain wholeness of character, a balanced relationship among desires, the victim of a dependency is incapable of reasonableness. If freedom is , impossible without rationality, this person cannot be free. But total ra-., tionality, absolute freedom, are unattainable. John Rawls's perfectly rational person in the original position, lacking in any defining or limiting habits or qualities, is avowedly a fiction of negative liberty. The corresponding ! fiction of positive liberty is the wholly integrated person in a society combining the greatest possible easy self-possession and self-control in each person with nearly frictionless cooperation among individuals. Any limiting habit seems incompatible with both kinds of freedom, but we all have patterns of behavior that limit us.

Mill said that anyone who could potentially be guided to improvement , by conviction or persuasion should be free of paternalistic coercion. But this may be impossible for alcoholics or heroin addicts, even though their reason is sound. Samuel Taylor Coleridge described alcoholics and opium addicts (including himself ) as suffering from "idiocy and lunacy of the ,! will" a strange notion, since idiocy and lunacy normally prevent the formation of a genuine will. A remarkable feature of drug abuse is that it calls out such extreme comparisons merely by intensifying the paradox of habit. Yet Coleridge's phrase remains a metaphor; the reasons for denying the retarded and insane certain rights and excusing them from ! responsibility cannot apply where reason is intact and the victims are apparently suffering only from their own deliberate acts.

The problem is particularly difficult when criminal responsibility has to be determined, and in the hardest kind of case, the criminal act concerns drugs only — possession, use, sale, or intoxication. In Robinson v. California (1962), the Supreme Court considered a state law that made "using or being addicted to the use of " narcotics a crime. In an opinion delivered by Justice Potter Stewart, the majority declared that imprisonment for addiction would be cruel and unusual punishment under the Eighth Amendment, because addiction is a status rather than an act; it is more like syphilis or insanity than criminal behavior. Justice John Marshall Harlan, in a concurring opinion, said that he would have allowed the imprisonment of Robinson, even though he was an addict, if the state had introduced evidence that he had actually talcen narcotics in California, but the use of addiction as a way to circumvent restrictions on evidence was objectionable. In dissent, Justice Byron White said that Robinson was not being punished for a status but for habitual use of narcotics, a series of acts. The Robinson case seemed to establish addictive compulsion as a potential bar to conviction for drug possession or even being under the influence of drugs. But the case has been interpreted very narrowly, along the lines of Harlan's concurring opinion. If the state can show proof of narcotics use (which is almost always necessary for proof of addiction in any case), anyone can be convicted, addict or not.

In Powell v. Texas (1968), the Court decided that it was not cruel and unusual punishment to jail an alcoholic for public intoxication. Justice Thurgood Marshall, speaking for four members of the majority in a five-to-four decision, criticized the idea that alcoholism implies compulsion or loss of control. He called the Robinson case irrelevant, because being drunk in a public place was clearly an act rather than a status; the only question was whether the act should be regarded as freely chosen, and the Robinson decision did not imply that the drunkenness of an alcoholic or the narcotics use of an addict is involuntary. Justice Abraham Fortas, speaking for four dissenters, said that this alcoholic's public drunkenness was the symptom of a condition over which he had no control and therefore no responsibility. 'White agreed with the dissenters that people should not be subject to criminal penalties for acts not under their control, but he concurred with the majority because he thought that Powell's public drunkenness, as opposed to his alcohol abuse, was voluntary.

What could not be achieved by constitutional interpretation is now being partially achieved by legislation; more than half of the states have passed some version of the Uniform Alcoholism and Alcoholism Intoxication Treatment Act, which takes the handling of public drunkenness out of the criminal law system and transfers it to public health authorities (with uncertain results so far). But the Robinson decision is as far as the courts have been willing to go on their own. In U.S. v. Moore (1973), the Supreme Court rejected an addict's claim that he could not be convicted of possessing heroin; it said that even if his possession of heroin was a symptom of the disease of addiction (which the court did not concede), it could be regarded for criminal law purposes as voluntary. Where the charge is more serious — selling drugs, robbery, assault — a pattern of addiction or alcoholism has never been an excuse for denying responsibility (although actual intoxication may affect specific intent and therefore reduce the seriousness of the crime).

An insane person can plausibly be said to lack the mens of mens rea (guilty mind) — the coherent mind that in legal terms makes responsibility possible (Fingarette and Hasse 1972). Yet the insanity defense is much criticized and rarely successful. It is hardly surprising, then, that no court has dared to propose an addiction defense. Incapacity allowing exculpation has never been interpreted to include mere difficulty in not desiring some-thing, or any mere habit, however powerful and well entrenched. Whatever loss of self-control or internal compulsion may be involved in drug de-pendence, we do not consider it an excuse for criminal acts. Coleridge's metaphor of "lunacy of the will" has never been carried that far.

The Court 's decision in the Robinson case proved to be less important than a dictum in which it expressed approval of involuntary civil commitment instead of criminal penalties for addicts. Here the issue is the opposite of exculpation. May drug abusers be confined against their will, under , the rubric of treatment or civil commitment, for a longer time than they could legally be forced to spend in prison, or in circumstances where they could not legally be imprisoned at all? By 1962 most states already had civil commitment provisions for alcoholics and narcotics addicts, but these were not often used. Civil commitment in the United States was mainly restricted to narcotics offenders in the federal hospitals at Lexington, Kentucky, and Forth Worth, Texas. In the 1960s, New York and California established civil commitment programs for addicts convicted of drug possession or sale, but they were of dubious effectiveness. Given a choice between prison and "treatment," addicts tended to choose whichever alternative meant a shorter term of confinement.

The constitutional basis for civil commitment of addicts has been litigated very little. In civil commitment hearings, courts have generally not required the procedural safeguards of the criminal law — the privilege against self-incrimination, right to confront witnesses, right to counsel, proof beyond a reasonable doubt, and so on. The District of Columbia Circuit Court suggested in 1971 that if there was no genuine treatment, civil commitment might be a cruel and unusual punishment in violation of the Eighth Amendment. The court was not clear about what would constitute genuine treatment, although it implied that more than an honest effort would probably be necessary. Despite this suggestion, civil com-mitment for addicts will probably never be rejected on substantive con-stitutional grounds (Dershowitz 1973). So courts have never analyzed the theoretical questions about civil confinement in cases involving addiction as opposed to insanity. In general, there are two requirements — incapacity or incompetency and danger to self or others. Is drug dependence the kind of incapacity that justifies preventive confinement? What likelihood and degree of danger should be required? Should the standards be different when persons other than the drug user are endangered?

The ambiguities of determining incapacity for free choice in drug users are familiar by now. At times a desire seems so irrational, wrong, or unwanted, and dominates a life so much, that we are prepared to treat it as a sign of incapacity and try to give effect to what the person who feels it would want if he or she were whole. On the other hand, giving in to an unwanted, irrational, or inappropriate desire is succumbing to temptation, and our literary paradigm of succumbing to temptation, Adam and Eve in the Garden of Eden, is also a paradigm of the exercise of free will. If we do determine that someone has become incapable of deciding freely whether or not to use drugs, Mill's principle is apparently no longer necessary. If free action is impossible, the right to do to yourself as you please has no meaning, and coercion is permissible to prevent harm to someone who is already unfree. Therefore the usual formula in involuntary civil commitment for insanity is "dangerous to themselves or bthers" without distinction. And in cases involving addiction, courts have never specifically required that the confinement be for the addict's own good. (An odd contradiction might arise if courts were more explicit, since it could be said that one of the main benefits of civil commitment for addicts, who are thereby in effect judged not to be responsible for their drug use, is that it might prevent them from being repeatedly jailed for drug posses-sion — that is, held criminally responsible for the same drug use by the same legal system.)

It is interesting to contrast this situation to one in which courts have been unwilling to impose even a much milder deprivation of liberty, despite a much greater and more immediate danger to both self and others. People are allowed to refuse blood transfusions for religious reasons even when they are likely to die and children are dependent on their support. The basis is the guarantee of religious freedom in the Constitution; but behind that, as in the case of the Native American Church and its peyote use, lies an assumption that the refusal, however eccentric or even irrational it seems to most of us, is free, wholehearted, a desire of the undivided person, and therefore worthy of respect.

Existing law makes little formal use of the distinction between harm to self and harm to others. But even in a legal system that in theory does take harm to others more seriously, harm to self might be a better reason for involuntary preventive confinement simply because it is easier to predict (Brock 1980). Dangerousness in the mentally ill, for example, is notoriously hard to judge, but the chronic suffering produced by an illness like schizophrenia and the relief afforded by the available palliatives are fairly well understood. The important word here is "illness." The course of schizophrenia is more or less predictable because it is a disease. Courts rely on the same interpretation of narcotics abuse when they allow in-' voluntary confinement; it makes the harm to the drug user seem predictable and the idea of a treatment or cure plausible. But are they?

This brings us back to medicine. We saw how effective the medical 'or disease model has been in justifying limitations on access to drugs, and we discussed some medical conceptions of addiction and dependence. Drug abuse and drug dependence are used as diagnostic classifications :by psychiatrists because they are now regarded as being, in some sense, illnesses. Whatever we may think about the metaphysics of free will, surely an act that is a symptom or defining feature of a disease must have a special moral and political status. The question of illness and the question of freedom and responsibility are intimately related.

Alcoholism has been analyzed more than any other form of drug abuse, since it is the most widespread and most severe. By now there is general official agreement that it should be called a disease, yet many people remain stubbornly reluctant to accept that view, and many who do accept it are unsure of just what makes alcoholism a disease or what kind of disease it is. In detail, the picture is utter confusion, and it begins to look as though the convenient label "disease" disguises an almost total lack ` of reliable luiowledge or plausible explanations.

One problem is that "in the first place, it is difficult to sustain any categorical statements about alcoholism" (Fingarette 1970, p. 801). The Royal College of Psychiatrists defines it as a craving, a "narrowed drinking repertoire," drinking that dominates a person's life, and relapse after 'abstinence (Royal College 1979, p. 42). Mark Keller, writing in 1960 as editor-in-chief of the Quarterly Journal of Studies on Alcohol, refers to "implicative" (suspicion-provoking) or marked and repetitive drinking 'that produces bad social, economic, or health consequences for the drinker {Keller 1960). The same author, writing in 1972, emphasizes the inability to choose consistently when to drink and when to stop (Keller 1972). Another authority, Morris E. Chafetz, calls alcoholism a behavioral disorder manifested by a preoccupation with alcohol to the detriment of health, by loss of control, and by self-destructive behavior (Chafetz 1979). Some advocates of what is called the "distribution-of-consumption" theory of alcohol control prefer to emphasize the amount of alcohol consumed, which they say correlates very well with health problems (DeLint 1976).

The diagnostic manual of the American Psychiatric Association shows an interesting change. In its first edition, alcoholism and drug dependence were classified as personality disorders and used in a diagnosis only as a last resort, if the drinking or drugs could not be regarded as secondary or symptomatic (American Psychiatric Association 1953). The third edition of the manual classifies alcohol abuse (the term "alcoholism" is not used) and other drug problems as substance use disorders, a special category. They are identified by a pathological pattern of use (this is a substitute for "psychological dependence" in older formulas), impairment in social or occupational functioning, and a duration of at least one month. Many indications of pathology and impairment are suggested, but none of them is set off as definitive or even more important than the others. The evidence for a pathological pattern of alcohol use includes such things as intoxication throughout the day, inability to stop or cut down, binge drinking, blackouts, repeated temporary abstinence, and attempts to restrict drinking to certain times of the day. Impairment is indicated by violence while intoxicated, absence from work, legal difficulties, and arguments with family or friends. Alcohol dependence is defined as alcohol abuse with either tolerance or a withdrawal reaction. Personality disorder, especially antisocial personality, is mentioned as a predisposing condition (American Psychiatric Association 1980).

So, this disease has a peculiarly diverse set of symptoms. Everything from the quantity of alcohol drunk to deviation from the accepted drinking practices of a social group has been used to define it. One respected authority lists more than a hundred definitions and hypotheses (Jellinek 1960). Such popular tests as craving, loss of control, tolerance, and withdrawal reactions are inadequate — either empirically wrong, too vague, or tautological (Maisto and Schefft 1977). Tests based on drinking at certain times of the day or in certain amounts include too many people or exclude some whose drinking is obviously abusive. It has even been contended that the skid row or Bowery dweller, long the typical repre-sentative of alcoholism in the public mind, is often not a "true alcoholic" at all (Straus and McCarthy 1951). There is no alcoholic personality, no definite progressive course (Trice and Wahl 1962), and, of course, no generally accepted treatment; even for selected groups of alcoholics, there is no clear evidence that any one kind of treatment is better than any other (Armor et al. 1976). The goal of treatment too is unclear; certain (angrily disputed) studies suggest that abstinence is not the only answer (or "cure") for all people defined as alcoholic (Pattison 1976). Wildly contradictory judgments about the aims and effectiveness of treatment are a natural result of confusion and vagueness in the definition of what is being treated.

In fact, alcoholics appear to have nothing in common which differen-tiates them from the rest of us, except that at some time in their lives they are regarded by others or regard themselves as persistent users of alcohol in a way that is harmful to themselves, their families, or society. ln despair, some authorities have taken to using the plural "alcoholisms," suggesting that it is not one disease but many. The trouble is that there is no established way of classifying these alcoholisms and identifying 4 causes and treatments. Many authorities are coming to think of alcohol abuse as simply one common effect of a great variety of biological char-acteristics, personal problems, social conditions, and situations in life. That is certainly what the studies of abusive behavior by frustrated ex-perimental animals suggest.

Most of the basic ideas on this subject are old ones, systematized and '. elaborated but not much improved in the twentieth century. There has . alway been doubt about the proper moral description of excessive drinking. The idea of classifying it as a disease, introduced in the eighteenth century but not ratified by the American Medical Association until 1958, did not arise from any new scientific knowledge; we know little more in that sense than we knew two centuries or twenty centuries ago. The disease concept of alcoholism at first "appears and disappears like a will-of-the-wisp, leaving no lasting impression" (Howland and Howland 1978, p. 40). Before the eighteenth century, habitual drunkenness was regarded as a moral weakness and a cause of disease, especially liver failure and insanity, but not as an illness. Jonathan Edwards used the example of the drunkard to illustrate his view that desire and will are one; the drunlcard chooses his ruin, Edwards said. He gave no credence to the idea that the alcoholic is enslaved. But other Puritan divines did call habitual dninkenness a lcind of madness or "an incurable habit," which at least hints at illness. Thomas Trotter, in his "Essay on Drunkenness," published in 1804, called it a "disease of the mind" and said that mere preaching against it was useless. Benjamin Rush, in An Inquiry into the Effects of Ardent Spirits (1811), described the physical effects of alcohol abuse and called it a "derangement of the will." But he also treated it as a moral weakness; this mixture is still very common (Levine 1978). By the 1850s the disease concept was well established, not among doctors but among some tem-perance reformers who believed that alcohol was the cause and abstinence for everyone the cure. By the 1890s alcoholism, along with morphine and cocaine addiction, had been added to some lists of mental illnesses.

But the churches still had the strongest influence in defining alcohol problems, and their view was essentially a moral one. Doctors were poorly organized and had little voice. The great change came after the failure ' of Prohibition, as the medical profession began to assert itself more aggressively. A new development of the disease concept, propagated by Alcoholics Anonymous and by the Yale Center of Alcohol Studies, em-phasized individual susceptibility as opposed to the malignancy of alcohol itself. Since then the disease concept has become even more firmly es-tablished, although we now emphasize addiction or dependence less than abuse, which sounds suspiciously like bad behavior rather than illness.

Popular attitudes duplicate the confusion among the authorities. In one household survey (Linsky 1972), the main causes given for alcoholism were personality disorder, biological susceptibility, moral weakness, and, more rarely, social drinking and alcohol itself. The treatments most approved were medical attention, psychiatric care, willpower, religion, legal controls, education, change of spouse or job, and family pressure, in that order. In another survey, 58 percent described alcoholics as "sick," 37 percent described them as "weak," and the rest did not know how to label them (Boyd 1970). People have accepted the idea that alcoholism is a disease, and they are willing to turn alcoholics over to experts (although there are no established treatments except for the physical symptoms), but many of them also retain vestiges of the old moralistic conception. (Alcoholics too often regard their behavior as shameful; maybe that is why one of the symptoms alcoholism workers are taught to notice is denial.) It has been suggested that a theory of alcoholism should cover all societies and all levels of drinking, should be compatible with theories about other deviant behavior (pathological gambling, for example), and should be testable and useful in practice (Sargent 1976). No such theory is likely to appear. Maybe alcoholism, like creativity, is not susceptible to explanation by a testable theory, because it is a name for too many different things in different circumstances.

Let us examine what it means to say that alcoholism (and, by implication, other forms of drug abuse) is a disease. This is not simply a dispute about words, because the decision to classify something as an illness has many social consequences: It affects attitudes toward the alcohol abuser, the types of control considered appropriate, criminal responsibility during intoxication, and insurance payments. If alcohol abuse were considered to be not illness but "succumbing to temptation," as Thomas Szasz describes it (and as most people would have judged it before the nineteenth century), or a way of life, as others have claimed, we would have to think about these matters in a different way. E. M. Jellinek, one of the best-known American authorities, originally insisted that alcoholism im-plied loss of control, because he feared that too broad a definition — say, one based entirely on excessive use — would undermine the ethical basis of social sanctions against habitual drunkenness by identifying it as a symptom of illness. Later he decided that some types or stages of alcoholism were diseases and others were not; it was a disease only if the drinker had lost control or was unable to abstain (Robinson 1972). Apparently he, at least, thought it essential to make some clear distinction between misbehavior and illness. (The issue is further complicated by the need to avoid confusing alcoholism with the diseases it causes, such as cirrhosis of the liver and Korsakoff 's syndrome, or with the effects of tolerance and withdrawal, which resemble a disease in the obvious physical sense.)

By itself, the fact that drug abuse has a social component or consists largely of repeated actions is no reason to deny the possibility of classifying it as an illness. A disease is, among other things, an abnormality that prevents adequate social participation or functioning, and cure always implies some kind of social reintegration. And of course, people can be responsible for their own illness, whether it is heart disease or drug abuse. Some destructive behavior seems so much more serious than a mere problem of living or a bad habit that it earns the description "behavioral disorder." The American Psychiatric Association, for example, includes pathological gambling and kleptomania as well as substance use disorders among its diagnostic categories. Even what we regard as physical illness varies with time and place; it has been said that in the lower Mississippi Valley in the early nineteenth century, malaria was so common that it was seen not as a sickness, an abnormality, but as one of the inevitable burdens of life. Illness may not be simply what doctors say it is, but the definition of disease also depends partly on the state of medicine, and for that reason, too, is subject to historical change. The gluttons Dante plated in hell as sinners would now probably be regarded as suffering from the behavioral disorder of overeating or the disease of alcoholism. Sometimes we even classify criminal behavior as a symptom of disease by branding it the product of sociopathy or an antisocial personality, although we have not reached the point of denying that criminals are responsible for their crimes.

The categories involved can be called, colloquially, "sick" (including "crazy"), "bad," and "wealc." Despite the efforts of R. D. Laing, Szasz, and other antipsychiatrists, most people believe they can distinguish between madness and problems of living. And no matter how often we are told that criminals are antisocial personalities and therefore in some sense sick or crazy, we are likely to go on thinking of them as blameworthy. But the status of drug abusers and addicts has always been much more am-biguous. Twentieth-century institutions have tried to resolve the issue by classifying them officially as sick, but alcoholics are still often regarded, despite the apparent contradiction, as sinners against themselves (weak)and against others (bad) because of the same actions that are supposed to be symptoms of disease. A complication is presented by the fact that using alcohol, even in excess, is also a normal, quite respectable social activity. Some politicians have recently claimed alcoholism as an excuse (moral, if not legal) for taking bribes or visiting male prostitutes. They are trying to use the modem disease concept to excuse their actions while in effect avoiding the stigma of serious abnormality that is the usual price of denying moral responsibility. This kind of manipulation is presumably what Jellinek feared when he refused to describe merely excessive or abusive alcohol use as a disease. Admittedly, it could not be attempted with any other drug. No congressman in trouble would dare to introduce heroin, cocaine, or marihuana as an excuse, because using those drugs is almost as scandalous as extortion or homosexuality. And even in this case it did not succeed; the ambiguity that the politicians were trying to take advantage of worked against them in the end because it made for commonsense limitations in the public mind. But common sense can change, and the limits remain unclear.

In law the influence of the disease concept has been important, although indirect and usually inexplicit. One of the underlying questions in the Robinson and Powell cases was whether the defendant's criminal act was a symptom of disease, and whether, if so, that made him not responsible. The decisions did not formally depend on this issue, but they provided a rich source of metaphors and comparisons. The majority opinion in the Robinson case compared punishing the addicted defendant to jailing a leper for his disease. Harlan, in dissent, responded to this analbgy by pointing out that courts had allowed the confinement, even in prison, of typhoid and venereal disease carriers. Justice Tom Clark, in dissent, justified Robinson's conviction on the ground that the penalty was criminal only in appearance but civil in intent, since the purpose was to arrest the disease of narcotic addiction in its early stages. In his opinion in the Powell case, Marshall explicitly criticized the disease concept of alcoholism, and the dissenter Fortas in effect defended it, saying that Powell's public drunkermess was symptomatic of a condition he could not avoid — implicitly a disease.

Deciding that alcoholism or addiction is a disease does not settle the issue of whether Robinson's or Powell's acts were voluntary, since people must often be held responsible for acts they would not have performed if they had not been alcoholics or addicts. Justice White made this distinction in the Powell case: Powell's alcoholism may have been involuntary; his public cirunIcenness was not. But when the act is regarded as merely a symptom, it seems harsh to ascribe responsibility to the actor. And since alcohol and drug abuse are defined as misuse of certain substances, almost any harmful action of the alcoholic or drug abuser might potentially be regarded as a symptom. The close association between alcoholism, drug abuse, and antisocial personality, sometimes regarded in the medical and sociological literature as an empirical correlation, is partly a matter of definition; as the term "substance abuse" suggests, this disease not only causes but often actually consists in part of antisocial acts (Grinspoon and Bakalar 1978).

In criminal law the disease concept is a minor issue, since courts have rarely been prepared to deny the criminal responsibility of alcohol and drug abusers. A defense of addiction used against a charge of robbing a drugstore would get the same short shrift as a defense of pathological gambling in an embezzlement case; the criminal courts must draw a line, somewhere if they are to have any function at all. But in civil commitment cases, disease becomes a much more important issue. Addiction or alcoholism can be seen as a serious incapacity, rattier than merely bad or self-destructive behavior, because it is pictured as a disturbance of adaptive functioning, biological and social at once. It is treated by detoxification, a medical procedure that allows the alcoholic or addict to recover from the physical symptoms of withdrawal with the least possible danger and discomfort. Methadone maintenance for heroin addicts also seems to be a form of drug therapy. This creates a certain aura of medical expertise that makes it more plausible to regard civil commitment as treatment for illness and therefore obviously for the addict's good. Unfortunately, detoxification provides only temporary symptomatic relief; and substituting one opiate for another, as we do in methadone programs, is not treatment for a disease (much less a cure) but a way to reduce the inconvenience of addiction for the addict and society. Evaluations of drug treatment programs show that once detoxification or methadone maintenance has ended, we have no clear idea of what to do (Ogburn 1978).

Some evidence supports the embarrassing suggestion that untreated heroin addicts are more likely to stop using the drug than those who are treated (Waldorf and Biernacki 1982). But maybe the addicts who are treated are simply the more severe cases, so that even after treatment they are not as well as the ones with a milder illness. To judge whether that is so, we first have to decide what constitutes a severe case of addiction. Many addicts, maybe most of them, have made a mess of their lives. They are socially isolated, neglect their health, and have no job or family; they pass the time cotnmitting burglaries, selling dnigs on the street, waiting for a connection, and nodding off. Eventually, if they survive long enough, they are likely to end up in prison, civil confinement, a detoxification clinic, or a methadone maintenance program. Addicts who do not lead that kind of life often have enough fortitude and social support either to sustain the addiction on their own or to give it up on their own, untreated.

But this does not tell us what defines a severe case of addiction. In the purely medical sense, it means very little. Certainly it does not depend on the total dose. There is no evidence that the people who succeed in breaking the habit on their own are the ones who are least heavily addicted in a physiological sense. On the contrary, many of the soldiers who became addicted in Vietnam but immediately stopped taking heroin on their return to the United States had been using a drug that was purer and stronger than anything available to street addicts here. It is convenient to conclude that the least severely addicted persons are precisely those whose lives are not so wasteful and chaotic that they end up in the treatment system. But such reasoning is circular; under this definition, treatment can never be proved to have failed. Lacking an independent criterion of severity, we have to say that the evidence against the effectiveness of treatment is strong.

But if there is no effective treatment, civilly committed addicfs may be in confinement solely because of some possible future misuse of drugs. This possibility properly makes many people uneasy, especially since addicts are usually referred to civil commitment after an arrest for drug possession.

This combination of medical diagnosis with moral or legal judement may come perilously close to punishing people for drug use more or less as criminals, while denying them the procedural rights of criminal law and allowing sentences of indeterminate length on the ground that they are sick and should be confined for their own good until they get well. To civil libertarians, civil commitment of opiate addicts sometimes looks like a pilot program for that dubious utopia, the therapeutic state. There is no question that this disease-crime model of drug abuse allows great scope to the state by introducing disease concepts into criminal proceedings and identifying the presence of a disease in willed and often unregretted acts. As Nils Bejerot writes, "the addict generally does not suffer from his disease, he enjoys it . . . the patients . . . must be . . . kept free from drugs for a long period, with or against their will" (Bejerot 1970, p. xvii). It becomes part of the definition of this illness that the patients may have no right to decide whether they want treatment for it. The only familiar similar situation is insanity; we are reminded again of "lunacy of the will." The treatment need not even be for the drug user's own good if drug abuse is regarded as an epidemic; then Mill's principle becomes irrelevant, as we saw in Chapter 1. Harlan's analogy in the Robinson case between civil commitment of addicts and quarantine of typhoid carriers shows the effect of this idea. In thinking of drug abuse as an infectious disease, we assimilate the voluntary process of persuasion and example by which it spreads to the involuntary transmission of infection.

This allows us to regard drug users as somehow both helpless victims and free persons who must be blamed for their actions. U.S. public policy on opiates especially has tended to follow this pattern. After the Harrison Act outlawed all use of opiates except for legitimate medical purposes, the government questioned whether maintenance of an addict on narcotics was a legitimate medical purpose. By the early 1920s, a doctor could write of "the shallow pretense that drug addiction is a disease" (Prentice 1921, p. 1553), and the medical profession was abandoning addicts. That was consistent; the law punished addicts as willful criminals, and doctors should so regard them.

Today, opinion has shifted again, and addiction (now called "opiate dependence") is fully recognized as a disease, but addicts are still subject to west, and, except for those in methadone programs, are not allowed treatment for their most obvious physical symptoms, the withdrawal reac-tion. We deny medical treatment for these symptoms and then treat the acts that produced the symptoms as themselves evidence of illness, without ' renouncing our moral disapproval or removing criminal penalties. The oinbssteitrvu:odnbayl riessitdroarinetcs hoenincoercion built into procedures for dealing with criminals and the quite different ones built into procedures for dealing with sick people can both be evaded in this way. A curious example is

That the basic concem of compulsory therapy plans is suppressive rather than therapeutic and that ideologies of suppression and of therapy do not easily mix is most evident in , the qualifications that are generally introduced as to which addicts are to be eligible for T compulsory therapy as an altemative to jail sentences. . . . Thus, there is generally a Inuit set on the number of times a patient may avail himself of the therapeutic altemative, and a self-conunitted addict who seeks discharge before his time is up renders himself ineligible forever. It is as if one were to declare that an easily cured patient is sick, but a hard-to-cure patient is a scoundrel. (Chein et al. 1964, p. 333)

The most important restraint on our treatment of sick people is the 1 requirement that the patient feel ill and want to be cured; the most important restraint on our treatment of criminals is the requirement that they have . committed some harmful act. If the condition of addiction to certain drugs , or the habit of using them is an illness and a crime, both restraints may come to seem unnecessary. Since drug use is a disease, there is no need for evidence of a harmful act or safeguards for individual rights; since drug use is a series of criminal acts, there is no need for consent to treatment.

The case of In Re de la 0, decided by the Supreme Court of California in 1963, involved a man who was convicted of a narcotics misdemeanor and then given an indeterminate sentence of up to five years of civil confinement for addiction. The court rejected his contention that this penalty was criminal and therefore barred by the Robinson decision. He argued that he had been sentenced under the penal code, placed in the custody of the Department of Corrections, confined against his wish with visitation and mail restrictions, and subjected to the authority of a criminal parole board. The state pointed out that addicts who commit themselves voluntarily are subject to the same rules. The court said that despite the unfortunate use of criminal law terminology this place of confinement was not a prison; the petitioner was there for quarantine and treatment.

As medicine advances and society becomes more complex and ration-alized, more and more things come to be regarded as disease, and once a condition or form of behavior is classified as a disease, we rarely look at it again in any other way. We move toward the therapeutic state that identifies health with positive liberty (thus the epithet "sick sociéty"). Whether you approve of this tendency or disapprove of it depends on your views about freedom and about the institutional arrangements that define health and sickness. In any case, it alters the balance among the biological, personal, and social components that enter any conception of illness. The biological anchor for the disease of drug abuse is a chentical with definite effects on the body and brain. But even the change in emphasis from addiction or dependence to abuse is a movement away from biology. And now we have come to the point where one writer calls alcoholism a matter of "social learning" (Larkin 1979) and another, thinking about heroin use in the United States, says that "the natural history of drug addiction is like that of a society: it must be rewritten every few years" (Vaillant 1970, p. 497). It is hardly surprising that alcohol and drug abuse are the only common diagnoses associated with antisocial personality. Most heroin addicts, for example, are juvenile delinquents or petty criminals before they ever use heroin, and addicts who do not have this kind of history are the most easily "cured" (Bess et al. 1972). In the English clinic system for opiate addiction, the addicts are described not as patients but as clients — a designation they prefer (Judson 1975) — and the aim of the program is not mainly to make the addicts stop using the drug (as in the U.S. methadone maintenance system) but to find work for them and help them develop other interests. If those aims are accomplished, they tend to give up opiates.

The question is whether we should go on calling something a disease when its social component is so important that it can be said to lack a natural history in the ordinary medical sense and may at times be hardly separable from general antisocial behavior. From the personal rather than the social point of view, we can ask whether there is any other disease for which the best-known cure is a kind of religious conversion or meetings in which the sufferers confess their misdeeds and exhort one another to change their ways. The biological anchor seems to be lost.

Maybe art can help here. Drugs and alcohol are moving agents in a play many people believe to be the greatest ever written by an Ameri-can — Eugene O'Neill's autobiographical Long Day' s Journey into Night (O'Neill 1956). It is set in a summer house in Connecticut where the four Tyrones (O'Neills) are gathered together in the summer of 1910. The action takes up one day from morning until late at night, and it consists mainly of two revelations that are not really different — about illness and about the family itself. Mrs. Tyrone is a morphine addict; her actor husband James and two (adult) sons are heavy drinkers — alcoholics by most definitions. Early in the play, Mary Tyrone learns that her younger son, Edmund (Eugene), has tuberculosis, and the men come to realize with shame and despair that she has started talcing morphine again. Her addiction is presented, at the start, as a disease like tuberculosis; the sons accuse their stingy father of causing it by hiring a cheap, incompetent doctor who gave her too much morphine during childbirth. The husband in tum scolds her: "I've been a god-damned fool to believe in you." But he also admits that "it was a curse put on you without your knowing or willing it." She complains of the local doctor, "When you're in agony and half insane, he sits and holds your hand and delivers sermons about will power." She says, "I've never understood anything about it except that one day long ago I found I could no longer call my soul my own." According to Edmund, she has said that she hopes she will take an accidental overdose some day.

But things turn out to be more complicated. She tries to convince herself that she is taking the morphine for her rheumatism. At one point she says, "I've become such a liar . . . , especially to myself," and she calls herself a "lying dope fiend." She asks Edmund not to believe that she used his illness as an excuse to talce morphine again. And there are other suggestions that the addiction is purposeful. It enables her to achieve "a peculiar detachment in her voice and manner." James says, "Every day from now on there'll be the same drifting away from us." Under the drug's influence, she becomes "a ghost haunting the past," and Edmund says that she takes it "to get beyond our reach." Her memories go back to the happy days at a convent school, before her marriage, when she thought about becoming a nun. Edmund tells his father, "You've never given her anything that would help her to want to stay off it" — despite all the money he has spent on cures. Maybe it is a question of wanting, then, and the addiction is "a curse put on her without her knowing or willing it," just like the other misfortunes of her marriage — no more and no less. This is true to what we know of medical addiction; most people who are given morphine by doctors, even if they suffer a physical withdrawal reaction, do not go on craving and seeking the drug afterward.

Alcohol, too, is at times a sickness for the Tyrones. Edmund says of a poet with whom he feels an affinity, "Poor Dowson. Booze and con-sumption got him," implying that they are the same kind of thing. It transpires that Mary's father, too, died of drink and tuberculosis, and hereditary susceptibility is suggested. But more often the men's alcohol-ism is presented as a continuing problem of character and situation rather than something traceable to an outside cause such as a tuberculosis bacillus or a doctor's needle. When James criticizes Mary for her addiction, she accuses him of bringing up the older son, Jamie, to be an alcoholic by giving him whiskey to quiet him as a baby — implying an analogy with the morphine used on her when she gave birth. But this is presented' only as a dubious suggestion prompted by anger and malice. In general, the characters are much less concerned about who or what to blame for the drinking, which is thought of more as something the men do than as something that has happened to them. And beyond the alcohol and morphine there are other, vaguer addictions — James's penny pinching and his habit of compulsively accumulating real estate, which his sons regard as a degenerate form of their peasant ancestors' attachment to the land.
Drink and drugs move the plot, too, because whiskey and morphine allow the Tyrones to think and say things they would otherwise not think and say; all the revelations and self-revelations are booze and drugs talking. Drug and alcohol problems illustrate remarkably how hard it can be to distinguish the press of circumstance from the unfolding of character. It is hard to draw a line between what these people have done and what has been done to them. As the action ends, late at night, with Mary Tyrone in a narcotic trance and her husband and older son stuporous from whiskey, we feel that it is inadequate to describe the morphine addiction and alcoholism as clinically identifiable illnesses; instead they represent what the Tyrones have made of their life together, and the definition of a "cure" would be to become different people leading a different kind of life.

Not all drug dependence includes the personality in this way. Nicotine dependence involves a much more limited set of symptoms, and is more easily interpreted as a specific disorder in an otherwise psychologically and biologically normal person. Maybe drug dependencies form a continuum on this scale. But in general, the trouble with calling drug abuse an illness is that the standard of health is so hard to identify once the biological level at which all human beings have a great deal in common is left behind. The danger is exemplified by the practice of imptisoning political dissidents in psychiatric hospitals for "reformist delusions." This makes sense if the norm of mental health includes agreement with existing social and political arrangements. And so it should if existing arrangements are fundamentally correct and health is defined socially. That is the ultimate dream of positive liberty. But such confidence about what is socially healthy seems monstrous; a society is not an organism.

To identify something as a disease is to set it off from other social and personal problems. Schizophrenia is not an exaggerated form of eccen-tricity or maladjustment; it is a very special condition. But it has never been shown that an alcoholic is a special lcind of person with a special susceptibility, who drinks for reasons different from the ordinary social drinker's, in a different way, and with different effects. Distinguishing between alcoholics and social drinkers is not easy, and people slip back and forth between these categories (Miller 1979). In one study comparing people who described themselves as alcoholics with another group of heavy drinkers who did not describe themselves that way, the researchers found that the self-identified alcoholics were more likely to have been to meetings of Alcoholics Anonymous (AA) and more likely to believe in the need for abstinence, but were otherwise very similar to the control group.- They had more alcohol problems, but the difference was one of degree (Skinner et al. 1980). Similar patterns exist in other forms of drug abuse — even heroin dependence (Vaillant 1970). That is only another way of saying that the social component in the definition of drug abuse is great.

Defenders of the disease concept have little to say about two problems that exacerbate each other and confound the treatment of dnig and alcohol abusers: their unwillingness to assume the role of patient (sometimes called denial) and the medical community's reluctance to confer it. Here is what often happens:

. . .drug and alcohol abusers receive medical care of lower quality than that accorded to other patient populations . . . diagnosis is often delayed or missed, patients may be excluded from the health care system, refenals may be ineffectively made, if at all. . . .
Some of these deleterious effects for patients can be attributed to negative attitudes on the part of physicians and other health care staff. Institutional characteristics also contribute to the problem. . . . There are insufficient illness-categories for patients with alcohol, amphetamine, or other drug problems. These patients are often disruptive or diagnostically confusing. . . .
These patients are then assigned to a "management" category. . . . The shorter the time on the ward and the less staff time required, the more successful the case. (Chappel et al. 1979, pp. 253-4)

But ignorance, impatience, and hostility on the part of medical staff do not account for everything. In fact, these institutional and personal obstacles develop partly because health professionals realize, consciously or not, that they are being asked to solve a problem that is largely nonmedical in the context of their training. For similar reasons, health insurance programs often cover alcohol and drug abuse inadequately. We hesitate before taking on, in the name of curing illness, what may turn out to be an indefinitely protracted task of getting unhappy and trouble-some people to live better or behave better. Alcoholism researchers also complain that supervisors in industry are reluctant to accept the diagnosis of alcoholism in employees. But this reluctance may be caused by un-derstandable confusion. Diagnosis suggests a conunon, easily recognizable abnormal or diseased pattern of drinlcing, and it is often hard to extract any such pattern from the varying ways in which people use and abuse alcohol at different times in their lives.

The complaints of people who are called on to treat alcoholics can be quite touching:

. . . if we insist on total abstinence as the criterion, program results are not impressive. If we are more modest and reasonable, as we are with depression and diabetes, for example, then respectable improvement rates are not uncommon. Alcoholism may then be seen as a treatable illness rather than a revolving-door phenomenon, staff morale improves, patients become less fatalistic, better arguments can be made for funding purposes, and staff may be more easily attracted to the field. (Gottheil 1982)

Unfortunately, it is not just a historical accident that the idea of alcoholism as a treatable disease has been associated with the idea of abstinence as a treatment. No one doubts that depression and diabetes will continue to be regarded as illnesses no matter how imperfect the treatment. But alcohol abuse is so unstable as a disease entity that sometimes only the goal of abstinence seems to preserve its coherence. Without that goal, the disease theory of alcoholism might turn out to be a will-of-the-wisp after all.

The disease concept requires careful criticism because this, if anything, is the received view. But it would not have become so widely accepted if there had not been much to be said for it. Civil commitment of addicts convicted of possessing heroin is rarely as oppressive in practice as it seems in libertarian principle. Whether a term of civil confinement works as treatment or not, it rarely lasts longer than the alternative prison sentence, and addicts often prefer it. The rise of methadone maintenance has meant much less deprivation of liberty for addicts diverted from the criminal courts. The legal system is too conservative to draw out all the possible implications of the disease—crime model. And this model is, after all, very attractive; even police officers who are professionally trained to handle addicts as criminals do not object to their being assigned to metha-done maintenance programs instead of prison after conviction (McDonald 1973). The model has become popular precisely because of the way it blurs the distinction between treatment and social control (or improvement); this vagueness may be inherent in any plausible conception of drug abuse.

In a book called Models of Madness, Models of Medicine (1974), Miriam Siegler and Humphry Osmond make a good case for regarding alcoholism and drug abuse as medical problems. They discuss various ways of looking at madness (in the sense of chronic psychosis) and conclude — as almost all psychiatrists would agree — that it is best regarded as a true disease (or diseases) and not, for example, as a problem of living or an inchoate romantic rebellion against an oppressive society. Heroin dependence and alcoholism are a much more difficult issue. Siegler , and Osmond describe medical, psychosocial, and a number of what they call "moral" models. Admitting that all of these models have some validity and are used by most people at one time or another, they discuss the flaws in each one and conclude that the best one for most purposes is a mixture of medical and "restorative moral" models. The virtue of tljis approach is that they abandon the attempt to determine what drug abuse and alcoholism are in some unattainably objective, scientific sense. Instead, they concentrate on what kind of authority is best for dealing with drug abuse and what the drug abuser's responsibilities should be. Thus, they are more concerned with defining a medical or quasimedical model than a disease concept. They emphasize discontinuity rather than continuity between alcoholism and social drinking, because the passage from being a problem or nuisance drinker to being a patient is rarely gradual. Alcoholism is treated as a metabolic defect compounded by social and psychological problems. The "moral etiology" of this disease is irrelevant because it is of no use in treatment. Blaming society or the alcoholic's family leaves the alcoholic with no responsibilities and nothing to do. Reproaching the alcoholic for moral weakness is notoriously futile. In this medical—moral model, the alcoholic is not a bad person but can be a bad patient — one who does not fulfill the duty of seriously trying to get well. This way of conceiving the problem, they believe, is above all best for the alcoholic's morale and thus for the chances of improvement.

Siegler and Osmond are in effect clarifying, systematizing, and extending the views of AA. They believe that the successes of AA come from its definition of alcoholism as much as its treatment methods. AA regards alcoholism as both an illness (which can be arrested by abstinence) and a spiritual problem. It must be conceived as an illness because illness is ego alien: The formula is not "I am doing bad things with alcohol" but "Alcohol has gained power over me." Alcoholics are unfortunate, not evil, but their vulnerability is special, so they do not have to denounce liquor and all of their drinking friends. They do have the duties of making restitution to the people they have harmed and following the rules that will enable them to get well. And the decision to get well requires a spiritual crisis and renewal. Like the disease—crime model for treatment of heroin addicts, which it formally resembles, this model of alcoholism is somewhat incoherent logically; its incoherence reflects the medically and morally ambiguous nature of alcoholism. It also contradicts some of the evidence by making a sharp distinction between alcoholic and ordinary social drinking and by insisting rigidly on abstinence. But the model is humane and seems to work for many alcoholics.

Siegler and Osmond's approach to the analysis of medical authority is related to their views on alcoholism and addiction. They classify it into three types: scientific (biomedical), public health, and Aesculaiiian. The authority of scientific medicine is "sapiential," based on knowledge and expertise about physiology. The authority of public health medicine is supposed to have a scientific basis, but it is primarily moral and political, dependent on a social consensus. Aesculapian authority is the individual doctor's relation to the patient — sapiential, moral, and charismatic' at once. Scientific medicine, unfortunately, has so far contributed little to the understanding of drug and alcohol abuse. Public health medicine is the realm of drug education and laws controlling drug use and sales. It is the Aesculapian relationship between individual drug abusers and medical professionals who care for them that Siegler and Osmond most want to identify and defend. To them it is the heart of the medical model as applied to drug abuse and alcoholism; the mixed source of its authority is most appropriate for a problem that evades categories.

For social reasons, Siegler and Osmond want to circumscribe medical authority carefully. They are somewhat wary of public health medicine's habit of using the language of health and illness for quasipolitical situations, and they insist on a discontinuity between drug use that is illness and drug use that is not. Their approach is most compatible with a political emphasis on negative as opposed to positive liberty. The most common justification for drug and alcohol prohibition, after all, is an intervention of public health medicine in the name of something very much like positive liberty. But the Aesculapian idea is not needed in all situations; for example, the disease concept of drug and alcohol abuse should not be used in judging criminal responsibility.

As Siegler and Osmond recognize, different conceptions of disease and different models of medicine are needed for different purposes. Doctors and other medical professionals have many roles, and there are many kinds of medical knowledge. The proportions of expertise, charisma, and moral authority in medicine vary with the situation, the illness, and the relationship between doctor and patient. Especially when dealing with drug abuse and alcoholism, we are always at the border between free will and compulsion, between illness and other forms of suffering. Szasz to the contrary, conceptual precision and logical fanaticism are wrong here. We are not able (and may never be able) to classify such situations into exclusive categories using a positivistic conception of disease and a simple, rigid notion of freedom. The "sick" label has its dangers, but especially in contrast to the "bad" label, it has the virtue of not narrowing the range of solutions too much. To use the label rigidly therefore is to repudiate its purpose.

Reflection on the nature of chronic drug problems has become much more sophisticated; for example, the latest formulations of the American Psychiatric Association on substance abuse are a great improvement over earlier ones. But what we have come to realize is mainly how much we cjo not know. References to dependence and addiction are no longer so promiscuous because the inadequacies of these terms are better understood. For lack of an alternative, the disease concept is widely used, but not without recognition that it is precarious and has some questionable social implications. We have probably gone about as far as we can in analyzing concepts and bringing commonsense knowledge to bear on them. Further developments in the field may depend on an unpredictable scientific break-through. Meanwhile, history and sociology illuminate some areas not reached by either political theory or medical knowledge.

* "The junk [heroin] merchant does not sell his product to the consumer, he sells the consumer to the product. He does not improve and simplify his merchandise, he deg-rades and simplifies the consumer" (William Burroughs).