A high correlation between fear and analgesia classically conditioned to footshock in rats has been reported in the literature. However, it has never been directly tested whether or not fear is in fact causal to the production of conditioned analgesia. We therefore tested whether conditioned analgesia could be elicited in the absence of fear by employing two independent methods of fear suppression. First, areas of the periaqueductal gray (PAG) previously implicated in fear were selectively lesioned. Lesions of the dorsolateral PAG significantly attenuated conditioned analgesia and markedly decreased fear responses. Second, fear was attenuated via administration of chlordiazepoxide (CDP). Rats which had been conditioned while in the presence of CDP showed no reduction in conditioned pain inhibition. These results demonstrate that: (1) fear is not causal to classically conditioned analgesia and (2) the anatomical substrates for fear and conditioned analgesia are distinct but partially overlapping. The fact that fear is not a critical antecedent for classically conditioned analgesia suggests that classical conditioning techniques may be applied clinically to increase the effectiveness of some analgesic manipulations.