Eight experiments were conducted to investigate the phenomenon of spontaneous recovery, or memory improvement over time without repeated testing. While this phenomenon has been previously studied within the verbal learning tradition, evidence for its existence has been inconclusive. Experiments 1a through 5 demonstrated the reliability of the effect. The procedure was gradually modified throughout these experiments, leading to the following conclusions. First, retroactive interference (which is a necessary condition for recovery) is maximized when subjects are led to believe that their knowledge of the target list will interfere with interpolated learning. Also, assessing spontaneous recovery in a within-, rather than a between-subjects paradigm, is a more powerful approach, and allows for a more sensitive test of the phenomenon. Most importantly, spontaneous recovery was reliably produced. The proposed explanation for the phenomenon involves retrieval inhibition, and its subsequent dissipation, as the processes underlying recovery. More specifically, inhibition prevents subjects from generating target items on immediate tests. As the inhibition dissipates, more items can be recalled on later tests; this occurs despite the presumption that normal forgetting is also operating upon the target list.
Experiment 6 attempted to extend spontaneous recovery to an implicit, word-stem completion test. Following study conditions roughly similar to those used in the prior experiments, there was no evidence for either retroactive interference or spontaneous recovery on the implicit test. This demonstrates, at the very least, a study manipulation that dissociates explicit free recall with implicit word-stem completion. More interestingly, it suggests that retrieval inhibition might only operate upon intentional uses of retrieval, although more data would be required to confirm this hypothesis.
Experiment 7 applied the spontaneous recovery paradigm to directed forgetting. If subjects are using retrieval inhibition to "block out" to-be-forgotten items, then these items should recover over time. Results provide limited evidence for this conclusion. Retrieval inhibition, and its subsequent dissipation, are hypothesized to be the primary processes underlying directed forgetting.