The ethical debate about drug policy has centered on the justification of prohibiting any form of drug sale or consumption; it has been dominated by broader issues about the type of harms the state may legitimately prohibit and the appropriate uses of the criminal sanction. While this debate has not been sterile, it has largely been a forum for profound arrd possibly irreconcilable differences about the priority of liberty and the role of the state. Differences in underlying political convictions have been exacerbated by disagreement about the addictive and criminogenic properties of various drugs.

Recently, however, there have been signs of an overlapping consensus for liberalization on the part of all but the most committed prohibitionists. The key factor in the emergence of this consensus is a perception that virtually all values threatened by the use and sale of drugs are threatened more severely by their criminalization. Many Who do not oppose criminal enforcement in principle have come to question it in practice, because of its failure to prevent rampant drug abuse and its role in creating a deadly illegal market, eroding civil liberties, and corrupting public officials.

The consensus that has emerged, however, is a fairly narrow one, favoring a reduced emphasis on the criminal law as a means of controlling drug use, a deliberate reduction in enforcement, and a decriminalization of simple possession. In this paper, I want to address two sets of moral and jurisprudential issues raised by these compromise policies. The first set concerns the ethics of partial liberalization itself: how fair and consistent with the norms of Anglo-American jurisprudence are policies of punishing the sale but not the possession of drugs and of deliberate under- or non-enforcement of the laws against sale and possession? These issues are peculiar to a policy of partial legalization, and the dilemmas they pose might be seen by proponents of full legalization as revealing the incoherence or instability of half-way measures.

The second set of issues, on the other hand, involve problems that a policy of partial decriminalization shares with a regime of full legalization: whether to prohibit the performance of certain high-risk activities "under the influence" of specific drugs, and whether, or when, to treat the influence of drugs as an aggravating or mitigating factor in the commission of other crimes. Even the most optimistic proponents of full legalization concede that the removal of all criminal sanctions will contribute to an increase in some kinds of injuries and some kinds of criminal activities; given our experience with the renewed legalization of alcohol, the increase is likely to be significant, at least in the short term. If laws governing the sale and consumption of alcohol provide a model for legalization, they also suggest the difficulties in integrating drug sale, possession and use into a more libertarian jurisprudence.

User Immunity and Statutory Rape

The decriminalization of drug possession and use would merely increase the disparities that already exist in the legal treatment of possession and sale. All jurisdictions have different penalties for the use and the sale of controlled substances. Some have further increased the disparity in treatment between users and sellers, decriminalizing "soft" drugs like marijuana or reducing their possession to a regulatory offense, subject only to fine. Some pursue a de facto policy of not prosecuting users of harder drugs who are not suspected of commercial sales. While these measures may reflect a reasonable use of limited criminal-justice resources and an understandable compassion for addicts, they may do little to alleviate the conditions which aggravate addiction: the high price of the drug and the domination of the drug market by violent criminal organizations. The de jure decriminalization of drug use also places the law in the awkward position of permitting acts (purchases) which, except for homegrown drugs, require forbidden acts (sales) for their completion.

It is instructive to compare the moral, legal and practical balance struck by user immunity with the structurally similar regulation of sexual conduct with minors through laws against "statutory" rape:

Is there a non-paternalistic basis for punishing only one party?

Statutory rape laws are avowedly paternalistic, in the sense that they deny the voluntariness and legal adequacy of the minor's "consent" to the sexual act. Criminalization of drug selling alone could rest on a similar basis, with the drug user deemed incapable of consent to a transaction driven by his addiction or dependency. But it may rest on other grounds as well. John Kleinig has argued that we need not act paternalistically in criminalizing only the exploitation of addictions, even if the exploited party can be said to have consented to the transaction — we act not to protect the exploited party, thereby denying her capacity to consent, but to express disapproval of the exploiter.1 As Joel Feinberg has pointed out, Kleinig's rationale would justify the punishment of drug sellers but not of gratuitous distributors.2 This distinction is already reflected in the penal law of those states which recognize an "agency" defense to drug sale in cases where the defendant has merely procured drugs as a favor rather than sold them at a profit. (Of course, under present law, the defendant is still liable for possession).

Is the distinction between the exempt and liable classes defensible?

Many confused and vulnerable older adolescents are technically guilty of statutory rape; many addicts sell just to maintain their habits. Many "victims" of statutory rape are at least as mature and sophisticated as the "perpetrators;" a smaller proportion of drug users are likely to be savvy consumers. Some states grade statutory rape by the disparity in the ages of the participants as well as the absolute age of the complainant; some jurisdictions exempt or downgrade small scale addict sales. As in many areas of criminal law, there is a tension between the need for clear, simple lines and for boundaries that reflect the actual contours of culpability, which may be quite convoluted. In the areas of adolescent sexual activity and drug transactions, the enormous variations in power, knowledge and motivation among the participants make the legislative task particularly difficult.

Does the distinction serve its intended purpose of protecting one class against the conduct of another?

Statutory rape laws have undoubtedly increased the bargaining power of the junior partners in many adolescent relationships, but their deterrent effect on precocious sexual activity is uncertain. User immunity may actually worsen the addict's situation in several ways, e.g, by raising drug prices to reflect the disparity in risk between seller and buyer. While it may make dealers less suspicious of user-informing, for which there will be less incentive, it may make them more suspicious of user-blackmail, for which there will be less deterrent. It may also cause the further erosion of the line between sellers and users, as pressure to equalize liability causes drug selling to "pyramid." In any event, it would not be surprising to discover effects quite different, and less benign, than those intended.

Non-enforcement and Under-enforcement as Deliberate Drug Policies

It would clearly be less controversial politically to cease or reduce the enforcement of drug laws rather than to explicitly permit the use and sale of drugs by legislative action. De facto decriminalization or under-enforcement would also provide a means of discouraging the upsurge in drug use that might follow outright legalization, by maintaining a moral stigma against drug use and the formal threat of arrest and prosecution. A policy of complete nonenforcement would not threaten unfair treatment to individual "offenders," but might undermine the integrity and moral authority of the criminal law. By deliberately maintaining criminal laws it has no intention of enforcing, the state undermines the publicity principle, divorcing the stated and the actual rules of conduct established by a legal system.

Even more acute problems are raised by a policy of under-enforcement, driven by regulatory concerns rather than scarce resources. Under such a policy, individuals would be punished for conduct that was not regarded as inherently criminal, as a means of controlling the undesirable effects of widespread activity. Laws intended to be enforced sparingly present many of the same moral problems as laws made to be broken, which have been the subject of considerable philosophic and jurisprudential commentary.

Thus, consider the non-paternalistic reasons that would favor a policy of occasional enforcement or nonenforcement over complete legalization. One principal concern is what Joel Feinberg calls the "garrison threshold," described by Douglas Husak as follows:

At some point, conduct that would initially be categorized as private and self-regarding may become so pervasive and widespread that it poses a threat to society. Any community can tolerate self-regarding conduct that is harmful to the agent unless and until it is practiced by a significant portion of the population, and crosses the threshold into serious public harm.3

Even if we had no moral objection whatsoever to drug use by adults, we might find that widespread use by adults led to large increases in street crime and in drug use by children, that such increases could not be prevented by stricter enforcement, but that they could be reduced, though not eliminated, by occasional crackdowns on adult sale and use. While this seems to provide good reason for keeping the drug laws on the books and enforcing them with an occasional sweep, it raises a moral problem about those arrested an prosecuted to keep the level of activity below the "garrison threshold."

The moral problem with such selective enforcement can be seen by distinguishing three types of partial enforcement: 1) against a random sample of those engaged in an indisputably criminal acts, with selectivity compelled by scarce resources 2) against a random sample of those engaged in an activity, like low-level pollution, whose marginal effects are slight but whose cumulative effects are severe, with enforcement set to keep the activity and the resulting harm at a tolerable level; 3) the same as 2), except that the activity causes no harm below a certain threshold, and the level of enforcement is set to keep the activity below that threshold.4

In 1), the target of selective enforcement cannot complain that others were not arrested and prosecuted as well, unless the grounds for selecting him were discriminatory. By committing an indisputably harmful act, he enters an enforcement lottery, whose outcome will be fair as long as its selection procedure is not biased. In 2), the target of selective enforcement has likewise committed a harmful act, although the harm may be de minimis if viewed in isolation. While he may deserve less punishment than the defendant in 1), he is no less deserving of punishment. The harm in 1) may depend as much as the harm in 2) on what other people do (consider a man who pushes someone in front of a moving subway); the difference is in the magnitude of harm directly attributable to the defendant.

In 3), the target's acts will cause no harm at all unless a certain number of other people act in the same way. The target arrested to keep the number below that threshold has done nothing wrong, unless he can be regarded as contributing probabilistically to the harm, by increasing the odds that the threshold will be exceeded. The justification for punishing him seems even weaker when, as in the case of drug activity, the feared harm involves the intervening criminal acts of other people. If fuel emissions become toxic when they exceed a certain atmospheric ppm, each emitting driver directly contributes to the harmful result. In contrast, the conscientious drug seller contributes to the (risk of) street crime or drug use by children in a very attenuated way if neither he nor those in his "chain of distribution" commit street crimes or give drugs to children. Thus, a policy of occasional enforcement, designed to keep drug activity below a critical social threshold, may be as indefensible morally as it is tempting politically.

Is the Criminalization of Conduct Done Under the Influence of Drugs Consistent with a Legalization Policy?

Alcohol regulation would appear to show the consistency of legalizing the use of a drug while criminalizing its combination with other activities. But the policy of punishing drivers based on their level of intoxication rather than their performance on the road departs from ordinary criminal jurisprudence, which justifies punishment only for conduct that causes harm or creates a specific risk of harm. The offense of driving while intoxicated (DWI) is not directed at reckless driving, which is separately proscribed, but the mere propensity to drive recklessly that is induced by alcohol. Modern DWI statutes do not even require physical impairment, as measured by a field sobriety test, but prohibit driving with a blood alcohol content (BAC) of .10 or higher. As James Jacobs argues, DWI laws rest on a conclusive presumption of recklessness; one that cannot be rebutted by showing a high tolerance for alcohol or an ability to drive safely under its influence.5

The treatment of elevated BAC as per se criminal, rather than as presumptive evidence of recklessness in the operation of a motor vehicle, punishes skillful but inebriated drivers whose performance creates less of a risk than that of unskilled but sober drivers, liable only when their actual driving violates standards of safe operation. DWI laws thus share with laws against drug possession a preemptive approach to social harm: both are based on a belief that intoxication leads to dangerous conduct, and on a policy of intervening to prevent that conduct from occurring.

Moreover, the focus on intoxication as a source of dangerous conduct, though based on some statistical evidence, also reflects disapproval or condemnation. As Jacobs has argued:

[Wihile drunk driving is undoubtedly dangerous behavior, more than a concern about dangerousness drives the law on this subject. Drunk driving is also an offense against morality, decorum and good citizenship. To understand why there is a special body of law aimed at the drunk driver, the offense must be seen in the context of social attitudes towards alcohol and alcohol abuse.... Historically, drunkenness was viewed as a moral failing. Driving while drunk continues to evoke the image of a social miscreant.6

Unless the criminalization of DWI can be justified by the administrative convenience of using drug-level as a surrogate for incompetence or recklessness, it seems to treat drug consumption as an evil above and beyond the objective risk it creates.

An alternative policy would require that all drivers maintain a minimum level of alertness, perceptual discrimination and motor skill. That minimum competence would be subject to testing if the police observed irregular driving, and perhaps also at random "competence" checkpoints. Obviously, the use of some drugs would impair the driver's competence, making it more likely that she would be subject to testing and would fail if tested. But a competence standard would come closer than a physiological standard to maintaining the requirement of harmful or risky behavior, and would treat drugs no differently than any other factor that might impair the driver's performance, such as sleeplessness or anger.

Some states already have laws against drugged driving; it is likely that the legalization of controlled substances will lead to a proliferation of laws criminalizing various activities done under their influence. These laws raise the same problem of consistency as DWI laws, suggesting a disapproval of drug use and a rear-guard action against the restriction of criminal liability to conduct causing actual harm and very specific risks. But they also raise distinct issues of enforcement and offense definition.

On the one hand, the legalization of drugs will faciiitate the enforcement of DWI laws, since drivers will be less covert about their chemical intoxicants, and perhaps less discrete about displaying symptoms of their use. On the other hand, the varieties of drug impairment are so great that the diagnostic value of the standard field test, designed to detect alcohol-induced impairment, will be severely limited. Unless we have a comprehensive driving competence test that can be reliably administered at the roadside, we will have to rely on either different competence tests for different drugs, or on per se blood content standards. The latter might become feasible with advancing pharmacological knowledge, at least of standard drugs. But as the number of types, brands and special effects proliferate, we may face a choice between overbroad "zero tolerance" for drugged drivers, or drug-specific maximum levels whose constant revision may cause acute problems of fair notice.

Is the Use of Drug Consumption as a Mitigating or Aggravating Factor in Criminal Offenses Consistent with a Legalization Policy?

The full or partial decriminalization of drugs will renew a perennial debate in criminal jurisprudence, over the role of, and proper response to, self-induced intoxication in criminal activity. It is undisputed that a high proportion of violent crimes are committed under the influence of alcohol and other drugs. It is also clear that their influence often makes the offender less responsible for his criminal conduct at the time he commits it, by distorting his perception of the circumstances and the character of his actions, and by reducing his scruples, fear and capacity for self-control. Moreover, these very features of intoxication are deliberately exploited by some would-be offenders to overcome their inhibitions. Because drug use both facilitates and extenuates criminal conduct, it creates a tension between the need to control violent crime and the duty to punish offenders in accordance with their culpability.

Proponents of decriminalization may incline to opposite sides of the debate. Those who see drug abuse as a medical or psychiatric problem may also be inclined to take a less punitive view of the crimes that it contributes to; those who favor decriminalization as a way of enhancing individual autonomy may be inclined to hold drug users fully responsible for their conduct while intoxicated. It is instructive to review the approaches taken in recent criminal law and jurisprudence, to examine their consis tency with a policy of decriminalization.

There are really two distinct issues raised by the use or exclusion of alcohol as a criminal defense: first, whether the consumption of a drug can prevent the defendant from forming the mental state required for culpability and second, whether, or how, the intentional, reckless or negligent use of a drug having this effect should affect the defendant's liability. Both issues have been addressed at length in the criminal jurisprudence of the past two decades. The first is controversial but highly-drug specific: psychologists and psycho-pharmacologists disagree about the extent to which different dosages of particular drugs affect perception, belief, emotion and self-control. While their disagreement may turn on differences about the clinical effects of the drugs, it may also reflect basic moral differences about free-will and responsibility. Most of the moral and jurisprudential debate, though, concerns the second issue: the legal response to self-induced intoxication conceded to preclude the mental state required for the offense.

Three approaches to intoxication as a defense can be discerned in modern criminal law. The first, found ip the Model Penal Code, simply treats intoxication as one factor relevant to determining whether the "mental elements" of the offense have been satisfied: e.g., if the charged offense requires "the intent to permanently deprive the owner of property" or "the awareness of a substantial risk of physical injury," the jury determines if the defendant's level of intoxication precludes (or raises a reasonable doubt as to) that intent or awareness. While this approach is consistent with the jurisprudence of modern penal law, it threatens to reduce or eliminate liability for a large proportion of the violent crimes committed by heavily intoxicated offenders.

A second approach avoids this consequence through legal devices which preserve at least partial liability for the most serious and common offenses committed under the influence. One such device permits intoxication to negate only the "specific intent" required for aggravated forms of an offense, such as "assault with intent to commit serious physical injury," not the "general intent" required for the assault itself. This allows intoxication to reduce the grade of the crime but not to acquit the defendant completely. The distinction on which it rests is highly artificial, since the intent required by the basic forms of assault and rape, however "general," will sometimes be negated by intoxication, e.g., if the defendant is too intoxicated to realize that he is striking a live object.

A second device, adopted by the Model Penal Code as an exception to its general "elements" approach, bars intoxication as a defense to crimes that require only recklessness or negligence, even if the defendant is too intoxicated to recognize any risk in his subsequent conduct. While the defendant unable to form "an intent to permanently deprive" as a result of intoxication would enjoy a defense, the defendant oblivious to a "substantial risk of injury" for the same reason would not. Clearly, this distinction between intentional and reckless/negligent crimes is as arbitrary as that between specific and general intent. The best that can be said of these distinctions is that they achieve rough justice, at the cost of considerable uncertainty and confusion.7

The third approach attempts to steer a middle course between the first two: while it exempts the intoxicated defendant from liability for any offense he lacks the mental state to commit, it imposes liability for the voluntary intoxication which prevents him from forming that mental state. One version of this approach is contained in a provision of the German penal code discussed by Fletcher, which imposes a flat punishment on any person held not responsible for criminal conduct committed while intentionally or negligently intoxicated. According to Fletcher, the defendant must be negligent not only as to the risk of becoming intoxicated, but the risk of committing a crime while intoxicated.8

It is instructive to contrast this provision with the offense of DWI, which is also designed to punish conduct creating a high risk of later misconduct. Unlike DWI, which does not require reckless or otherwise illegal conduct by the intoxicated driver, the German provision imposes liability only if the risked conduct occurs. As Fletcher notes, this suggests that the provision is not directed at negligent risk-creation so much as at partnership or complicity in the subsequent offense: the person who gets intoxicated, intending or strongly suspecting that he will act violently as a result, is treated like a responsible adult who lures or incites a juvenile into committing a crime for which the juvenile is not legally responsible.9

But if the defendant can be regarded as an accomplice in his own subsequent misconduct, why not punish him more severely for aiding and abetting more serious misconduct? A person who drinks himself into a violent rage with the intention or expectation of killing his lover is clearly more culpable than a person who drinks to acquire the "dutch courage" for a car theft. If the defendant can be reckless or negligent as to the generic risk of committing a crime, he can certainly be reckless or negligent as to the risk of committing a specific crime. Like the American restrictions on the intoxication defense, the German provision on crime-excusing intoxication works a very rough compromise between the conflicting imperatives to deter and to extenuate intoxicated criminal conduct. By imposing the same penalty regardless of the magnitude of the resulting crime, it still treats intoxication as a separate evil, rather than as an integral part of the offense it facilitates.

A recent proposal by Paul Robinson refines the German approach, scaling the defendant's liability to the offense he extenuates by intoxication. Robinson would allow intoxication to negate any element of the offense, but would hold the defendant liable if he satisfied that element "immediately preceding or during the time that he was becoming intoxicated ... and the harm or evil he intended, contemplated, or risked is brought about by [his] subsequent conduct during intoxication."10 Thus, if the defendant takes a drug he knows will make him uncontrollably violent, intending to kill his lover, he will be liable for intentional homicide, even if the actual killing is excused as involuntary. If he does not intend to kill his lover but merely disregards a high risk that he would become uncontrollably violent in her presence, he would be guilty of reckless homicide, not for the killing itself, but for recklessly disregarding the risk of its occurrence.

This proposal develops Fletcher's suggestion that the intoxicated defendant be treated as an accomplice to his own misconduct. It tracks the conditions of the standard complicity statute, making the defendant liable for a later offense if he contributes to its occurrence with the mental culpability required for the offense itself. Like the acts of encouragement and support that make a person liable as an accomplice — acts which may be perfectly lawful in isolation — intoxication is not treated as criminal in itself, but only as facilitating a specific criminal offense.

Robinson's proposal thus comes closest of those considered to assimilating intoxication to the general principles of culpability and excuse in contemporary jurisprudence. It treats intoxication neither as a separate evil nor as a source of blanket immunity. In contrast to the American compromise approaches, it does not draw artificial lines between the offenses for which intoxication can and cannot be a defense. Unlike the German approach, it scales the defendant's culpability for promoting his own criminal conduct to the seriousness of the offense he intends, expects or risks. Its complexity may be unavoidable, given the complex relationships between intoxication, crime and culpability.

A general issue faced by all three approaches is determining when, if ever, self-induced intoxication is voluntary. Those who use alcohol or drugs most abusively are also likely to use them most compulsively. In assessing the voluntariness of habitual intoxication, researchers and policy-makers differ on both the strength of the drug habit and the responsibility of the user for acquiring it. Again, their disagreement reflects a complex mix of empirical and moral differences.

The aggravating role of alcohol or drugs in criminal law is clearly inconsistent with a decriminalization policy. Even if the consumption of certain drugs were shown to increase the frequency of crime, or violent crime, there would be no basis for punishing more harshly the commission of such crimes under their influence. The use of drug-consumption as an aggravating factor seems to rest on an inapt analogy to the use of a weapon, which increases the objective risk of additional harm in the course of the offense. To justify the similar treatment of drug-consumption, it would be necessary to show that the drug in question not only increased the incidence of the offense, but the probability of further harm during its commission. Even then, it would be problematic to treat the formation of a psycho-pharmacological state as equivalent to the possession of a weapon.

Conclusion

The case for liberalizing the drug laws may be simple and compelling, but the ethical and legal implications of almost any liberalization policy are complex and controversial. Different forms of liberalization raise distinct concerns: in decriminalizing possession but not sale, we greatly increase the disparities in the legal treatment of the two parties to a drug transaction; in using infrequent enforcement to keep drug use to a tolerable level, we punish people for conduct we do not regard as criminal. Other concerns are raised by all forms of liberalization: in removing legal deterrents to drug use, we must develop appropriate legal responses to the almost certain increase in crimes and high-risk activities performed under the influence of drugs. These issues are as important as questions about the social and economic impact of liberalization, and need to be as thoroughly debated. I hope this brief review has succeeded in calling attention to their importance and complexity.

Footnotes

1. John Kleinig, "Consent as a Defense in Criminal Law," Archives for Philosophy of Law and Social Philosophy (Weisbaden: Franz Steiner Verlag GMBH, 1979): 340-344.

2. Joel Feinberg, "Non-coercive Exploitation," in Rolf Sartorius, ed., Paternalism (Minneapolis: University of Minnesota Press, 1979): 227.

3. Douglas Husak, "Recreational Drugs and Paternalism," Law and Philosophy 8: 353, 362.

4. The distinction between 2) and 3) is loosely derived from Derek Parfit's discussion, in "What We Do Together" (unpublished ms., 3/29/89), of the spectrum of joint action: from threshold effects, where one person's acts makes no difference if a certain number have (already) acted and fully addictive effects, where each person's act makes an independent contribution.

5. James Jacobs, Drunk Driving: An American Dilemma (Chicago: University of Chicago Press, 1989): 61-62.

6. Jacobs, Drunk Driving: An American Dilemma, 64.

7. See George Fletcher's critique of these compromise measures in Rethinking Criminal Law (Boston: Little, Brown, 1978): 846- 852, and Jacob's discussion in Drunk Driving: An American Dilemma, 62-63.

8. Fletcher, Rethinking Criminal Law, 847-848.

9. Id.

10. Paul Robinson, "Causing the Conditions of One's Own Defense: A Study in the Limits of Theory in Criminal Law Doctrine," Virginia Law Review 71: 1, 61 (1985).