Semantic associates create retroactive interference on an independent recall task

Semantic similarity between stimuli can cause false memories, but the extent to which it causes retroactive interference in recall has been less explored. Here, subjects learned unique locations for “critical” words that reliably produce false memories in the Deese-Roediger-McDermott paradigm. Next, subjects centrally viewed words that were semantically associated with half of the critical words. Finally, subjects retrieved the critical word locations (to test recall) and distinguished them from previously unseen words (to test recognition). We found spatial memory impairments for critical words whose semantic associates were shown (vs. not shown), suggesting that semantic material caused retroactive interference, even on a test of unrelated content (i.e., spatial versus semantic). This effect was present in three experiments when the interfering information was presented shortly before spatial recall, but not after a one-hour delay between associate learning and test or after swapping the order of the spatial and associate phases. Moreover, impairments occurred whether or not subjects were aware of the semantic relatedness between critical and associate words and consistently occurred when the associates had low-to-moderate strength in predicting the critical words. By contrast, swapping the order of the two learning phases increased critical word recognition in a manner that scaled linearly with associate-to-critical word strength. These findings suggest that memory impairments can occur solely via semantic associates on an independent task where all relevant responses are freely available; in this way, they cannot be attributed to any conventional account of retroactive interference.