Some of the considerations that led to a consolidation interpretation of retrograde amnesia (RA), which states that RA results from the disruption of memory processing and storage when neural activity is interrupted by a brain insult, are reviewed here. The time-dependent gradient of memory loss (i.e., new memories are more vulnerable to amnesia than old memories) that characterizes RA seemed to fit nicely with the notion of a cascade of cellular events occurring during the immediate post-acquisition period that would transform a labile representation into a more stable form (i.e., consolidate the memory). However, a variety of observations came to challenge the storage-disruption model, and among these was the finding of amnesia for old but reactivated memories. A recent study by Nader, Schafe, and LeDoux (2000) provides an important analytic extension of the work on "reconsolidation" by showing that inhibition of protein synthesis in the lateral and basal nuclei of the amygdala immediately following the reactivation of old memory will induce retrograde amnesia. We offer a retrieval-oriented conceptualization to account for the temporal gradient and the "reconsolidation" phenomena.