ABSTRACT
This chapter examines different instances of updating and knowledge-revision from the integrative perspective of an extended construction-integration model, which involves the component processes of activation, integration, and validation of information. It discusses the available empirical evidence for when and how readers update or revise existing representations, and when they fail to do so. This discussion will prepare the ground for an outline of basic principles and cognitive processes that contribute to discourse updating and revision of existing knowledge structures. Validation may be regarded as a precondition for revising existing knowledge structures in long-term memory. Long-term working memories (LT-WM) bridge the gap between limited capacity of working memory and the need for efficient access to prior discourse ideas. In the language domain, LT-WMs comprise the situation model of the discourse up to the segment. Sperber et al. proposed that language users possess a basic capacity called epistemic vigilance that capitalizes on a protective function of validation when acquiring knowledge from texts.
