Four experiments examined the effect of naloxone pretreatment on the expression and extinction of ethanol-induced conditioned place preference (experiments 1, 2, 4) or conditioned place aversion (experiments 1, 3). DBA/2 J mice received four pairings of a distinctive tactile (floor) stimulus (CS) with injection of ethanol (2 g/kg) given either immediately before or after 5-min exposure to the CS. A different stimulus was paired with injection of saline. Pre-CS injection of ethanol produced conditioned place preference, whereas post-CS injection of ethanol produced conditioned place aversion. Both behaviors extinguished partially during repeated choice testing after vehicle injection. Naloxone (10 mg/kg) had little effect on the initial expression of conditioned place preference, but facilitated its extinction. Moreover, repeated naloxone testing resulted in the expression of a weak conditioned place aversion to the CS that initially elicited a place preference. In contrast, naloxone (1.5 or 10 mg/kg) enhanced expression of conditioned place aversion, thereby increasing its resistance to extinction. A control experiment (experiment 4) indicated that repeated testing with a different aversive drug, lithium chloride, did not affect rate of extinction or produce an aversion to the CS previously paired with ethanol. These findings do not support the suggestion that naloxone facilitates the general processes that underlie extinction of associative learning. Also, these data are not readily explained by the conditioning of place aversion at the time of testing. Rather, naloxone's effects appear to reflect a selective influence on maintenance of ethanol's conditioned rewarding effect, an effect that may be mediated by release of endogenous opioids. Overall, these findings encourage further consideration of the use of opiate antagonists in the treatment of alcoholism.