An extension of Merton's formulation is the theory of Cloward and Ohlin (1960), with its focus on the structure of opportunities. They posit a more elaborate organization or criminal activity, in which youngsters who are recruited into crime achieve some of the culturally prescribed rewards associated with achievement. But those who have failed in both the conventional route and the criminal one are double failures and prime candidates for drug use. The significance of this formulation is that it also locates heroin use among the structures that are external to the individual. It appears to comport with the fact that minority youths, who are assumed to have little access to mobility in the ranks of organized crime, have higher rates of drug use than do lower class white youths, who presumably have such access.

It is difficult to document the distribution of various forms of organized crime, or the recruitment of youngsters into these circles, except in illustrative or anecdotal ways (White 1943). The body of findings we will review later suggests that addicts are derived from the same matrix found in nonaddicted delinquents and that there is little to distinguish them from nonusers. And although addicts generally commit fewer violent crimes than nonaddicted criminals, they must be quite good at various forms of hustling and criminal activity in order to survive (Lukoff 1972; Preble and Miller 1977; NIDA 1976).

Despite their failure to explain drug use, the theories of Merton and of Cloward and Ohlin continue to be influential. The problems in specifying universal norms, or reasonably coherent structures that allocate individuals along different paths, are not unique to these formulations. But they are the major efforts that have as their goals the identification of socially structured alternatives within which individuals presumably act out their lives and shape the options available to them (Stinchcombe 1975). While contingent and subcultural patterns may contribute to different modes of expression, they still attempt to specify the broad outlines that direct persons' lives.

Much research on drug use would appear to examine derivative themes that are useful for organizing much of our knowledge. We make no effort to review all of the research in this brief paper, only that portion that directs us to alternative structural sources for understanding drug use and deviance.

SOCIAL LOCATION

Most investigations, even those that are descriptive or primarily epidemiological, without any clear theoretical agenda, generally examine substance use rates by age, sex, social class, and race/ethnicity (Abelson et al. 1977; Johnston et al. 1979; O'Donnell et al. 1976). Social class and race/ethnicity serve as surrogates for socially significant structural parameters. Where patterned differences emerge, they appear to reflect the different propensities these groups have for drug experimentation. The theoretical issue, at first glance, is to comprehend how social location affects individuals located differentially within society.

But illicit substance use is very volatile, even over relatively short historical epochs. Currently, heroin use is concentrated in black and Hispanic communities, which appears to suggest that both lower socioeconomic status and belonging to disadvantaged minorities provide

important clues to the attraction of heroin use. However, at the turn of the century, opiate use, in various forms, was found primarily in white, middle-class females as a result of therapeutic use (Ball 1970; Ball and Bates 1970; Commission of Inquiry into the Non-Medical Use of Drugs 1973). In Britain, heroin users roughly match the class distribution of the larger society, and blacks are underrepresented (Commission of Inquiry into the Non-Medical Use of Drugs 1973). A closer examination of heroin use in ghetto communities in the United States reveals a more complex relationship of stratification to heroin use. Vaillant (1966b) contrasted Lexington addicts against their own communities and observed they were better educated than their comparable age-mates in the same tracts. In a survey of an urban ghetto community, it was found that reported heroin use was associated with higher socioeconomic status, although this, as we will see, was a spurious relationship (Lukoff and Brook 1974; Lukoff 1977). Thus, heroin users are not necessarily drawn from the most impoverished segments of the communities, where use is currently concentrated (Nurco 1979; Robins 1975a). Nor, as their education and intelligence suggest, are they necessarily those who should appear to be doomed to the margins of society (Ball and Bates 1970). Only when contrasted against the larger society do socioeconomic status and lower education appear to be related to heroin use. This, however, appears to be the wrong way to examine the information. Instead, the relevant contrast would appear to be to examine heroin users against the backdrop of their own communities. Then the picture shifts substantially.

Because heroin use is a relatively rare event, most general population surveys report too few users for reliable estimates. Thus, caution is necessary in interpreting trends. In a study of selective service registrants, O'Donnell and his colleagues (1976), when examining reported narcotic use by cohorts, showed that there was a decline among blacks in the later cohort, with an accompanying increase among whites. In a survey of blacks in Harlem, Brunswick and Boyle (1979) examined rates by cohorts and observed a decline in initiation into heroin use among younger members of their panel. Although it bears repeating that caution should be used, such trends do suggest how ephemeral heroin or other narcotic use might be in historic perspective, and that the clues to its use might be elsewhere than in the simple matter of gross contrasts by class or race observable in any one epoch.

The dynamic nature of drug use trends is even clearer for marijuana. When Becker (1963) investigated marijuana use two decades ago, it was largely confined to inner city blacks and jazz musicians. Currently, marijuana competes with alcohol as the most popular drug, especially among the young (Jessor and Jessor 1977; Johnston et al. 1978; Kandel 1978a). Jessor and Jessor (1978), in reviewing marijuana trends, observe that there is a declining significance of such factors as "urbanicity," race, and socioeconomic status. Even sex differences are declining, although they appear to persist for heroin use. "At the level of the demographic environment then there has been a trend toward homogenization as far as variation in marijuana use is concerned" (Jessor and Jessor 1978, p. 341). It is increasingly smoked in public settings; legal penalties in many places have been reduced; sanctions, where they exist, are often not invoked for possession of small quantities for personal use. Even when sanctions were punitive, marijuana use continued to increase in popularity, both for those who have ever tried it and among the proportion who use it with reasonable frequency. Thus, normative systems are often only marginally effective, and they

are subject to rapid change as the larger community begins to accommodate the persistent and pervasive use of the substance.

If most of the usual indicators of social location show declining significance, one persistent feature of marijuana use continues to be important: The vast majority of users are young. And increasingly, the age of onset of marijuana use appears to be declining (Abelson et al. 1977; Johnston et al. 1979). Because of the relatively short time in which marijuana use has become popular, it is possible that current youthful and young adult users will continue to use it as they become older.

The same persistent relationship to age is present among heroin users. Almost all users start when young, at least in the United States experience (Brunswick and Boyle 1979; Lukoff 1972; Nurco 1979; Robins 1975a). As cohorts of adults advance in age, the largest proportion who were addicted abandon heroin use. Winick (1964) estimates the typical duration of addiction to be just over eight years. Although older addicts exist, the heroin-using population is still weighted toward those who are relatively young.

Thus, the one unambiguous association with drug use, one that appears to persist, at least in Western cultures, is the relationship of drug use to youthfulness (Braucht et al. 1973). Most of those who experiment with illicit drugs are young; those who become addicted, where there is information, decrease or cease drug use with advancing age. Structural variables such as social class and race/ethnicity are much more ambiguously related to drug use, as our review of trends suggests. We exclude the misuse of medically prescribed drugs because they would appear to present very different configurations.

SOCIALIZATION

The identification of social norms assumes that behavior is transmitted to actors who, depending on circumstances, tend to adhere to appropriate beliefs and concomitant behaviors. This explains the emphasis on socialization in the research literature, although sometimes only the "end product," the beliefs themselves, is identified and assumed to have been somehow transmitted (Jessor et al. 1968; Merton 1957). The search for antecedents of personality, rooted in family childrearing practices, overlaps with the effort to identify how cultural values and norms are communicated to the young (Brook et al. 1977a,b, 1978; Lukoff 1977).

But socialization is not limited to the family. Other agencies of social control also contribute, sometimes with perspectives that are at variance with those of the family. The most heavily investigated area has been the impact of peer groups (Becker 1963; Braucht et al. 1973; Feldman 1968) and the attendant mechanisms that shape the choice of friends and influence the accommodation to the behaviors and values of peers. This raises two theoretical issues. The first is the identification of the countervailing forces that influence the decline of parental legitimacy, as well as of other agencies that promote conventional behavior. The second issue is the way in which adolescents develop a peer culture with alternate value systems and goals (Becker 1963; Feldman 1968; O'Donnell et al. 1976; Whyte 1943).

There is one thing, however, which is not altogether congruent with the above statement. The literature on family socialization of adolescents has two foci, and many variations within each. First, there is a focus on the models family members provide for the use of drugs, tobacco, alcohol, or even medically prescribed, mood-altering drugs (Brook et al. 1977a, 1978; Kandel et al. 1978). Here, the assumption is that children will emulate their parents' use regardless of the choice of substance. From this perspective, although substances may change, there should be a continuity across generations. The findings are generally consistent with this assumption, although less powerful than one might expect. This may be an artifact because rates of reported use by family members whether obtained from adolescents or from parents are generally very low compared to the rates of usage of illicit drugs by adolescents. Alcohol, of course, differs in this respect from illicit substances (Braucht et al. 1973). The direct modeling of parents' behaviors is unlikely to explain a great deal of the usage by younger individuals where rates of use decline rapidly after the mid-twenties (Abelson et al. 1977).

The other focus is the examination of various forms of childrearing as well as the quality of the parent-child relationship, i.e., whether there is warmth and affection between them. These studies generally indicate that parental rules are related to adolescent drug use (Brook et al. 1977a, 1978). More proscriptive orientations are associated with lower rates of drug use. In addition, adolescents who report positively on their parents also tend to have lower rates of involvement with illicit substances (Gerstein 1976). This is an important research direction that has its own utility.

From the perspective laid out at the beginning of this paper, however, these are intervening processes. Since the purpose of this paper is to identify aspects of the social structure that ultimately affect adolescents and young adults, it is necessary to recast the issue in order to attempt to understand what it is about the social structure that may result in variations in the form of socialization.

Although not ordinarily viewed in the context of socialization, age of onset of drug use serves as a surrogate index for an important dimension of socialization, namely, the unfettering of the bonds of social control. Early onset reflects the premature segmentalizing, or insulation, of youthful activities from the normative system of the adult community. Those who start young are more likely to persist in substance use and other forms of deviance and to resist the blandishments of treatment (Commission of Inquiry into the Non-Medical Use of Drugs 1973; Lukoff 1972; NIDA 1976). Referring to narcotics users, Nurco (1979, p. 321) states, "The earlier the onset of deviant behaviors, the more malignant the process invoked and the more ominous the prognosis. . . . The younger the age of onset the more intense and committed the addictive career."

The early onset of drug use and other forms of deviance means that individuals are less likely to complete school, to have a history of sustained employment, or to engage in other adolescent or young adult activities that facilitate passage to adult status. In this sense, their socialization is truncated, and they are less prepared to assume the requirements of adult roles of their communities. They are only marginally connected to the adult worlds of their respective communities. Adolescent lifestyles, congregating with peers, avoiding employment

and family relationships may persist until the person is quite advanced in age (Preble and Miller 1977).

Equally important is the fact that, as Robins (1979) has noted, any. form of deviance, particularly among the young, forecasts other forms of deviance, including alcohol consumption, school deportment, delinquency, and early sexual promiscuity. There are several possible implications, but the one that concerns us here is that the roots of deviance are shared by many forms of problem behavior. The form that problem behavior takes, while it may in part reflect personal dispositions, is primarily a response to the encounters with other individuals, the peer cultures of adolescents and young adults.

The longitudinal reconstruction of substance use by Robins (1975a), from premilitary usage through Vietnam and after discharge, dramatically illustrated how easy access to heroin inflated use rates substantially. Almost all soldiers in Vietnam would presumably have had easy access to heroin, but not all of them used it. But those who scored high on preservice deviance were about four times as likely to initiate use as those who were low in deviance. These findings underscore that while proximal settings, where drugs are plentiful, markedly affect rates of use, earlier histories also exert a powerful influence. The fact that among heroin users there is often a history of delinquency prior to the onset of use is consistent with these findings (Lukoff 1972; NIDA 1976).

Despite the addictive potential of heroin, for some individuals involvement is only experimental or sporadic; others appear to cease use without the assistance of treatment or to accommodate the goals of treatment programs. Although scarcely studied, the information that is available indicates that such individuals are less alienated or disengaged from family and work, and less intensively immersed in drugusing groups (Lukoff 1974; Robins 1979; Zinberg 1979).

Because marijuana is a common recreational drug for so many persons, the factors involved in its use are more diverse than those for heroin use. It appears necessary, for example, to distinguish between persons in a late-onset, sporadic-use group and persons in an early-onset, frequent-use (generally daily) group. For those in the first group, use is confined to specific social contexts in which it is simply a cultural trend, much like tastes in music or clothing (i.e., the use is governed by proximal variables reflecting aspects of the current social milieu). In the second group, use can be predicted from antecedent variables, such as perceived or actual parental roles and the quality of familial relationships (Jessor and Jessor 1977; Jessor et al. 1968; Braucht et al. 1973; Brook et al. 1977a, 1978; Lukoff 1977).

As marijuana use moved from vanguard users who adopted the drug when it was still subject to heavy penalties, it appears to have also attracted individuals who, in varying degrees, were less likely to be engaged in subcultures that held perspectives divergent from those of the larger society.

Although concepts used in the many investigations reflect the general anarchy in a great deal of social research, one trend appears to persist, namely, that youthful onset of marijuana use is associated with a slackening of parental controls, early rebelliousness, and the presence of a wide array of behaviors incongruent with the expectations of the family, i.e., adult controls are markedly attenuated so that the