When the Facts Just Don't Add Up: The Fractal Nature of Conversational Stories

The present study frames the patterning of factual, schematic inference, and interpretive information over time in conversational storytelling using dynamic systems theory. This approach is consistent with current conceptualizations of remembering, as well as with psycholinguistic approaches to storytelling in discourse but goes beyond extant work by presenting a novel way to quantify the temporal patterning of different kinds of information in stories. The patterning of facts, schematic inferences, and interpretations in stories from 104 dyads discussing a computer game experience was examined, and a quantitative index of the extent to which that patterning was temporally dependent was created. That index reflected aspects of current conceptions of coherence in narrative, as well as some additional aspects special to conversational stories. As hypothesized, the temporal patterning varied as a function of both experimentally varied and speaker-rated listener attentiveness. Moreover, the temporal patterning of information across all but one story in the sample suggested that standard summary approaches to stories will be limited in their ability to capture important features of stories. Implications are discussed in terms of this new approach in relation to current practices.