ABSTRACT
A core process involves updating, which people define as constructing, adding to, or revising existing representations of characters encoded and stored in memory. This chapter describes experimental research that has helped identify the ways in which readers use the products of updating processes to generate predictions and expectations for future narrative events. It discusses the situations in which readers learn about characters across multiple narrative episodes rather than in single events. When this process spans multiple texts, the cognitive demands may increase further, with readers potentially needing to recall and reconcile a broader range of representations developed over varying amounts of time. Readers' prior knowledge represents a rich resource that will necessarily be recruited to support ongoing understandings of events and descriptions provided across text experiences involving the same characters and settings. Readers spontaneously updated their expectations that Philip was smart only after reading a refutation with an explanation, as exemplified by slowdowns to sentences consistent with a potentially instantiated trait model.
