Recognition memory for Spanish-Catalan cognate and noncognate words was tested at retention intervals of 30 minutes, 3 days, and 7 days using a remember/know response procedure. We observed a clear mirror effect for the cognate-noncognate stimulus class and a remember-know response categorisation at the immediate retention interval. However, the cognate and noncognate mirror was still observed at 3 and 7 days, whereas the remember-know mirror disappeared at both retention intervals. Also, we ran a repeated testing condition to be able to carry out a sequential item analysis and observe the fate of the original remember and know responses 3 or 7 days later. The analysis supported the idea that there was a loss of contextual information that was at the root of the disappearance of the remember-know mirror effect. These results provide support to the idea that it is the imbalance between recollection and familiarity that is the most likely cause of the mirror effect.