Smokers' responses to smoking cues may be the result of a classical conditioning process. There is evidence that classical conditioning may not proceed in human subjects unless subjects are consciously aware of the stimulus (CS)-reinforcer (UCS) contingencies. In two experiments the role of CS<-->UCS contingency awareness in the expression of conditioned responses (craving, salivation, and skin conductance) was studied. A discriminative classical conditioning paradigm was used during which subjects were presented with one stimulus (the CS +) always paired with cigarette smoking (the UCS) and another (the CS -) never paired with cigarette smoking. Half of the subjects were given instructions to discover the CS<-->UCS contingencies (group 'aware', AWR), whereas the other half were not (group 'unaware', UWR). In experiment 1, all subjects responded to the CS + with increased cigarette craving relative to the CS -; this effect was more pronounced in the AWR group compared to the UWR group. A lower amount of salivation in response to the CS + compared to the CS - was found in the UWR group. These between-group differences were interpreted as a consequence of the enhanced expectancies of smoking in the presence of CS + in group AWR compared to group UWR. In experiment 2, the observed craving responses to CS + and CS - were consistent with those seen in experiment 1, but no discriminative salivary response to the stimuli was found. When, after conditioning training, subjects' expectancies of smoking were removed by instructions, and their responses to CS + and CS - were again measured, the discriminative craving response to CS + and CS - was eliminated and all subjects demonstrated a lower amount of salivation in response to the CS + compared to the CS -. These data suggest that presentation of arbitrary cues previously paired with cigarette smoking can elicit CRs and that facilitation of awareness of the CS<-->UCS contingency by instructions can potentiate craving CRs. In addition, these data suggest that craving CRs can be eliminated, whereas compensatory CRs can be facilitated, when cigarette expectancy is removed with instructions.