Readers who watch the American television show Dallus might recall the following episode: A public altercation at a major social reception at the Ewing ranch causes conflict, embarrassment and tension. JR Ewing, speaking to his father, Jock, sees a passing waitress with champagne and says something in the order of,“Here, you need one of these” as he reaches for a glass of bubbly. Jock replies that he needs “something a lot stronger than that.” The two go off to a local bar, get intoxicated, reappear happy, laughing and relaxed, and drive home intoxicated but safely. This vignette, aside from exemplifying the unfortunate modeling of a destructive drinking pattern (Wilson, in press), illustrates the widely shared conviction that alcohol reduces tension or anxiety. This, of course, is the well-known tension or anxiety reducing theory of alcohol use and abuse. It is not surprising that television would portray this function of alcohol. One of the most striking features of the anxiety reduction theory is how widely it is believed. Both social and problem drinkers subscribe to the belief, the latter even more strongly than the former (Edwards, 1972; Goldman, Brown and Christiansen, 1987). Therapists who treat alcoholism also hold to this view (Cappell, 1975). Even such erstwhile foes as psychoanalysis and learning theory agree on this putative property of ethanol. The superego has been defined as that part of the psyche that is soluble in alcohol. Learning theorists have looked to alcohol’s anxiety reducing effects as a negative reinforcer of maladaptive behavior. Not insignificantly, English literature and poetry contain some of the classic statements of alcohol’s anxiety reducing effects. Almost as striking as how commonly held this belief in alcohol’s anxiolytic effects has been, is first, the influence of the anxiety reduction theory of alcohol on the behavior of researchers across disciplines; and second, the difficulty we have ‘encountered in consistently demonstrating such effects in people in the experimental laboratory. Since it was formally spelled-out by Conger (1956), the tension or anxiety reducing theory has had a profound impact on research on alcohol. Three decades later it continues to generate studies and stir debate. The global terms “anxiety” and “tension” can be used roughly synonymously. Attempts to draw finer distinctions have not been successful. Stress reduction is another common and broadly comparable term. Most recently, Levenson and Sher have introduced the term “stress response dampening” to describe alcohol’s effects. Although it is narrower in scope, there is probably little to be accomplished by trying to distinguish the stress response dampening model from what has come to be known as the tension reduction theory (Blane and Leonard, 1987). Sher (1987) himself characterizes the stress response dampening model as “a modest, pared-down tension reduction hypothesis”(p. 234).

It is not difficult to understand why the anxiety reducing theory has had such broad and lasting appeal. Demonstrating robust anxiolytic effects of alcohol would help elucidate the causes of alcohol abuse and steer us towards what would need to be put right in order to help alcoholics. But the theory promised more than a plausible account of the development and maintenance of alcohol abuse. Potentially, alcohol’s commonly observed actions on a range of everyday activities might be explained. As Neal Miller, one of the pioneers of research in this area put it,“alcohol has a perplexing variety of effects, making some aggressive, others amorous, some tearful, and others talkative”(1964). However, if we assume that the diverse behavior that follows alcohol consumption …