LAUSR

Abstract Within-trial priming paradigms have been widely used to measure lexical retrieval and familiarity-based processes in speeded pronunciation, perceptual identification, lexical decision, lexical retrieval, and episodic recognition tasks. Here, we introduce a novel within-trial priming paradigm to examine cued recall, which is considered a more recollection-based task. In each experiment, participants initially studied a list of paired-associates (e.g., BADGE-gold). During each recall test trial, 500 ms after the onset of the stimulus cue (BADGE-??????), a pattern-masked prime was briefly presented for either 48 ms or 125 ms, with the task to recall the word (e.g., gold) paired with the cue (BADGE). Across seven experiments, the primes were identical (gold), semantically related (silver), orthographically related (good), or unrelated (chair) to the to-be-recalled response item (gold). The results consistently indicated that the masked identity primes benefited immediate recall at both the 48 ms and 125 ms durations. There was no benefit from the orthographically related prime condition, suggesting that participants were not using partial letter information as an additional cue for memory retrieval. Semantically related primes only produced a benefit in immediate recall at the 125 ms prime duration. Delayed cued recall performance indicated that the facilitatory priming effects observed in immediate recall were eliminated. In addition, results from conditional analyses suggest that the benefits of retrieval in an immediate recall condition are reversed in the delayed recall condition. We consider multiple interpretations of these delayed recall results and argue that the results are in part due to priming influencing an activation-based mechanism in the immediate recall task, which decreases the retrieval demands for the response item to the paired associate cue, and hence, decreases the long-term benefits of retrieval practice.