The contextualization of input and output events in memory

Several observations from everyday life suggest that people are deficient in monitoring their own actions, often forgetting that they have already performed a planned act, or experiencing doubt as to whether they have done so. These observations appear inconsistent with the many laboratory studies that indicate that people are quite efficient in monitoring their own actions. Towards the resolution of this discrepancy we proposed that: (a) output monitoring in real life often requires the retrieval of a specific, contextually framed episode rather than mere familiarity with an event, and (b) output events are less strongly integrated with their environmental contexts than input events are. Therefore, despite the output advantage that is frequently reported in occurrence memory, context memory should be relatively less efficient for output than for input events. This hypothesis received some support in Experiment 1, in which generated verbal responses were remembered better than read responses, but the difference was significantly smaller for context than for occurrence memory. Experiment 2 employed a task simulating a two-person interaction. While occurrence memory was superior for self-performed tasks to that for other-performed tasks, context memory was in fact inferior for the former tasks. These results were seen to suggest that self-initiated actions tend to undergo a weaker contextual integration than events that originate from a source external to the person.