In a typical Pavlovian conditioning experiment, a relatively insignificant event, the conditioned stimulus (CS), is paired with a biologically more meaningful event, the unconditioned stimulus (US). As a consequence of those pairings, the CS is thought to acquire response characteristics of the US. In this article I describe experiments with rats that suggest that under some circumstances, the CS acquires control of perceptual processing of the US, in the absence of that US itself. I present three kinds of evidence for this surrogate processing, which I liken to imagery or hallucination: (1) CSs come to control specific, sensory-evaluative responses normally evoked only by the USs; (2) CSs can substitute for USs in the establishment of new learning about those USs themselves; (3) CSs can substitute for USs in the modulation of conditioning to other events, either overshadowing (interfering) or potentiating learning, in the same manner as the USs themselves. Finally, I compare these data with evidence for conditioned sensation and imagery in humans, and suggest that imagery may be a very basic process, evolutionarily derived from perceptual and conditioning processes adapted to deal with remote or absent objects.