It has been widely reported – if not universally replicated – that presenting a reminder of a previously trained cue-drug or cue-fear memory before extinction training can lead to a long-term reduction in subsequent responding for the cue, that goes beyond the reduction in responding observed with extinction training alone. This ‘retrieval-extinction’ phenomenon has been hypothesised to depend upon memory reconsolidation; namely, that the cue-drug or cue-fear memory becomes unstable during the reminder session, and is overwritten by the subsequent extinction training. An alternative explanation is that retrieval-extinction leads to an enhancement of extinction. Here, I will present our recent research addressing whether retrieval-extinction is a reconsolidation-based or extinction-based phenomenon.