Stress, cues, and pharmacological priming are linked with relapse to addictive behavior. Increased salience and decreased inhibitory control are thought to mediate the effects of relapse-related stimuli. However, the functional relationship between these two processes is unclear. To address this issue, a modified Stop Signal Task was employed, which used Alcohol, Neutral, and Non-Words as Go stimuli, and lexical decision as the Go response. Subjects were 38 male problem drinkers (mean Alcohol Dependence Scale (ADS) score: 18.0). Uncontrollable noise (B10 min at 110 dB) was the stressor; nonalcoholic placebo beer (P-Beer) was the cue manipulation, and alcohol (0.7 g/kg), the pharmacological prime. Half the sample received alcohol, and half P-Beer. Stress and beverage (test drink vs soft drink) were manipulated within subjects on two sessions, with half the sample receiving active manipulations together and half receiving them separately. Go response time (RT) and Stop Signal RT (SSRT) were slower to Alcohol than Neutral words. Stress augmented this bias. Alcohol and P-Beer impaired overall SSRT. Stress impaired neither overall SSRT nor Go RT. SSRT to Neutral words and Non-Words correlated inversely with Go RT to Alcohol and Neutral words, and Non-Words. ADS correlated directly with SSRT to Alcohol words. A resource allocation account was proposed, whereby diversion of limited resources to salient cues effectively yoked otherwise independent Go and Stop processes. Disturbances of prefrontal norepinephrine and dopamine were cited as possibly accounting for these effects. Treatments that optimize prefrontal catecholamine transmission may deter relapse by reducing disinhibitory effects of salient eliciting stimuli.