The relationships among negative health behaviors and addiction are not limited to correlation among drug habits. Mechanic (1979) found smokers were less likely to wear seat belts, while Kalant and Kalant (1976) found users of both prescription and illicit amphetamines suffered more accidents, injuries, and untimely deaths. Smokers have 40% higher accident rates than nonsmokers (McGuire, 1972). From the standpoint of these data addiction is part of a panoply of self-destructive behaviors some people regularly engage in. Drunk drivers turn out to have more accidents and worse driving records than others even when they drive sober (Walker, 1986), suggesting that drunk driving is not an alcohol problem but one of drunk drivers' generally reckless and antisocial behavior. The disease model and behavioral theories both have missed the extent to which excessive and harmful substance use fits larger patterns in people's lives.

Drug Abuse as the Failure of Children to Develop Prosocial Values

The use of a combination of early-life factors to predict both heroin use and addiction to other drugs reinforces the results of a large (and growing) number of studies of adolescent drug use. Jessor and Jessor's (1977) pioneering work emphasized a kind of nonconfomity dimension in predicting both drug and sexual experimentation. This factor seems rather too global, in that it confuses personal adventurousness with antisocial alienation (not to dismiss the possibility that adolescents can confuse these things). Pandina and Scheul (1983) constructed a more refined psychosocial index on which drug and alcohol-abusing adolescents showed high scores, but on which "a large proportion of student moderate users did not display problematic or dysfunctional profiles" (p. 970). Further explorations in this area of research have indicated at least three interesting and potentially related dimensions associated with drug and alcohol abuse:

1. alienation. Adolescents who abuse a range of substances are more isolated from social networks of all kinds. At the same time (perhaps as a result), they associate with groups of heavy drug users that reject mainstream institutions and other involvements connected with career success and accomplishment (Kandel, 1984; Oetting and Beauvais, this issue). Individual orientations in part precede the selection of group association, although group involvement then exacerbates individual inclinations in this direction.
2. rejection of achievement values. Jessor and Jessor found that absence of achievement values strongly predicted drug use. In the Monitoring the Future study of the class of 1980, Clayton (1985) pointed out, second to marijuana use in predicting extent of cocaine involvement was truancy. Clayton speculated it was unlikely that cocaine involvement preceded truancy in these data, and thus the absence of a commitment to school attendance was a condition for drug abuse. Lang (1983) provided a summary of data indicating an inverse relationship between achievement values and substance abuse.
3. antisocial aggressiveness and acting out. A relationship between antisocial impulsiveness or aggressiveness and alcoholism has been repeatedly noted. MacAndrew (1981) reported 16 studies showing a higher (in some cases much higher) than eighty percent detection rate for clinical alcoholics through the MAC scale of the MMPI. The highest factor loading for the scale was "boldness," interpreted as "an assertive, aggressive, pleasure-seeking character," an example of "factor loadings that make alcoholics resemble criminals and delinquents" (MacAndrew, 1981:617). MacAndrew (1981) in addition noted five studies of clinical drug abusers that showed similarly high detection rates according to the MAC scale. MacAndrew (1986) has found a similar kind of antisocial thrill-seeking to characterize women alcoholics.

The MAC scale and similar measures are not measuring the consequences of alcohol and drug abuse. Hoffman et al. (l974) found the MAC scores for treated alcoholics were not significantly different from those the same subjects showed on entering college. Loper et al. (1973) also detected higher Pd and Ma scores on MMPI responses (indicators of sociopathy, defiance of authority, et al.) in college students who later became alcoholic. This finding is reinforced by similar results Jones (1968) obtained with young respondents through use of Q sorts.

These findings are so well established that the battle is to claim them for different domains of explanation. Genetic models of alcoholism now regularly incorporate the idea of the inheritance of impulsive, delinquent, and criminal tendencies. Tarter and Edwards (this volume), for example, postulated that impulsivity is the central element in inheritance of alcoholism. I have elsewhere summarized grounds for caution about such genetic models (Peele, 1986b). The crucial issue is the relationship between addiction as antisocial misbehavior and socialization processes and social values. Cahalan and Room (1974) found alcohol abuse was strongly related to antisocial acting out, but their data clearly identify this as a social phenomenon found among particular groups. The question I pose in this article is whether we see it as within our cultural control to minimize through social learning the expression of uninhibited aggression, sensation-seeking, and disregard for social consequences that characterize addiction.

The Commonplaceness of Natural Remission in Addiction

A crucial element in the disease myth of addiction, one used to justify expensive, long-term - and increasingly coercive and involuntary - treatment is the progressive and irreversible nature of addiction. According to one television advertisement, overcoming alcoholism on one's own is like operating on yourself. All data dispute this. Epidemiological research finds that people typically outgrow drinking problems, so that alcohol abuse decreases with age (Cahalan and Room, 1974). The data on drug abuse are identical, and less than one-third of men who have ever used heroin continue to do so throughout their twenties (O'Donnell et al., 1976). We have reviewed data such as Schachter's (1982) and Garn's (1985) which indicate that long-term weight loss is a common event. Yet perhaps the single greatest area of self-cure of addiction is smoking - approximately 30 million people have quit smoking, with ninety-five percent quitting on their own (USPHS, 1979).

Conventional wisdom about addiction denies this commonplace reality to such an extent that addiction and alcoholism experts often seem embarked on campaigns to attack their own data. For example, Vaillant (1983:284-285) combined data showing that a majority of alcohol abusers in his sample were in remission, hardly any due to treatment, and that his own hospital patients' outcomes after two and eight years "were no better than the natural history of the disorder" with an insistence that alcoholism be treated medically (Vaillant, 1983:20). Although he found the large majority of his natural-history population recovered from alcoholism without the assistance of AA (including even those who abstained), all of Vaillant's lengthy case studies indicated that this is impossible. (In further data from his study Vaillant has sent me, those who quit drinking by attending AA had higher relapse rates than those who quit on their own.)