Abstract

One of the characteristics of connected discourse is that entities introduced are often referred to again at a later point. Such anaphoric reference clearly necessitates a memory representation of the prior text, so that the new reference can be established as coreferential with the old. In this paper, a distinction is made between people introduced in a narrative who depend upon a particular setting in order to be interpreted (scenario-dependent entities), and those who do not (main characters). It is argued that the availability of representations of these two types of character in working memory will depend upon whether the text indicates a change in setting to have occurred. Two experiments are described which show that dependent entities become less available if a change has occurred, while main characters are not so affected. The first study uses incidence of mention in a passage continuation procedure as an index of availability, while the second uses reading time and question-answering latencies. The experiments rule out any simple view that availability depends only upon how far back in the text a character was mentioned, and illustrate how passage of time in a narrative serves as a cue for recognizing the end of an episode.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call