The Nature of Memory for Conversations: Questions and Organization

Abstract

The focus of this dissertation was to investigate how multiple factors influence the memory for conversations. First, we hypothesized that questions are less well remembered than answers. This hypothesis is based on proposals surrounding the functional role of questions as temporary placeholders in communication (Heller & Brown-Schmidt, 2023) and theories that posit the meaning of a question is defined by the possible answers (Groenendijk & Stokhof, 1984). In Experiments 1-2 we tested this hypothesis by having pairs of participants engage in communication games where they named objects in either question and answer pairs or assertions. We found that recognition memory was best for second named items in pairs, whether that was an answer or part of an assertion. Memory for answers may be better than memory for questions, but future research will need to disentangle the effect from order of presentation. Second, we hypothesized that temporal, speaker, and topic contexts would each influence the organization of conversation recall. In word list paradigms, temporal, semantic, and source topics have all been shown to influence the order of recall (Polyn et al., 2009). This study thus tested relevant contexts to see if similar principles guided conversation recall. In Experiment 3, pairs of participants engaged in 15 minute conversations about a list of six topics. They independently recalled these conversations after a 20 minute delay, a ~4 day delay, and a ~10 day delay. We found temporal contiguity at each delay. Participants were more temporally organized within self-produced idea units than partner-produced idea units, reflecting the memorial benefit for self-produced idea units in conversation. We find recall patterns to support the theory that there is a temporal context for topics and a separate temporal context for idea units within each topic. Across all three experiments, we find that elements of a given utterance, such as the role it plays in communication, the serial order, the speaker, and the content discussed all influence what is recalled and/or in what order it is recalled.