Both results and interpretations are conflicting concerning proactive or retroactive interferences that may be generated by individual items upon other items within a single recognition list. In an attempt to explore a factor that may assist in reconciling varied findings, each of the three present lists consisting of either CVCs, two‐digits, or words, respectively, was presented via half‐lists on multi‐study and ‐test cycles. Two half‐lists were given in such a way that both sublists experienced the same number of intervening events in the mean retention interval, but differing ones in the composition of mean retroactive study and test events. Sixty male and female undergraduates participated. In contrast to small proactive study event interference, both retroactive study and retroactive test events did control recognition performances in an intriguing manner: Relative interferences generated by retroactive study events were weaker early in acquisition, but gained strength as learning progressed, and at late acquisition stages became quite powerful vis‐à‐vis retroactive test events. Thus, the acquisition stage seems to be a relevant factor in determining recognition performance.