The role of affect in narrative structure

Publisher Website
Google Scholar
Abstract
Within the field of Artificial Intelligence, research on knowledge-based text analysis has shown how natural language processing depends on common-sense reasoning, plausible inference, and massive amounts of domain knowledge for even the simplest of children's stories. Working within this framework, we have identified a level of memory representation for narratives based on plot units. Plot units organise knowledge about emotional responses to stereotypical situations and human interactions. We claim that plot units capture the underlying structure of a narrative more effectively than analyses based on goal states or planning strategies. In this paper we describe the plot unit approach to text analysis, demonstrate a plot unit analysis for a short narrative, and discuss the relative merits of plot units vs. story grammars.