Abstract
It is becoming accepted that the associative strength of a cue can change in its absence. However, the rules that govern the association of the representations of absent cues are in dispute. There is evidence to support both the idea that the representation of a cue is associated in the same fashion as the cue itself (e.g. Holland, 1990) and the contrary idea that the representation will be associated in the opposite manner to the cue itself (e.g. Dickinson & Burke, 1996). In three experiments an aversion to almond was formed by the pairing of the induction of nausea by lithium chloride injection with a context previously paired with almond. In addition in the third experiment an aversion was formed to almond when a flavoured solution previously paired with quinine was drunk in a context previously paired with almond. These results suggest that the rules governing the association of the representations of absent cues depend on the type of reinforcer.