A Computational Model of On-line Story Understanding

Models of story understanding in artificial intelligence and cognitive psychology typically concentrate on the construction of representations of story content. Construction of a 'complete' representation (informally equivalent to an 'understanding') requires inferences which establish coherence links between text statements. This report examines how reality-based links (e.g. causal connections) and narrative-based links (e.g. those that establish rhetorical structure and delineate text segments) have been handled in previous models. My suggestion is that both types of links are required for narrative understanding, and that each is dependent on the other. I also indicate how the inferences required to create both types of link may be uniformly represented and processed within a story grammar framework. This framework could also be used to model theories about recall (hierarchical storage) and processing.