ABSTRACT
Text understanding involves more than the construction and retrieval of a mental representation of the explicit text. Rather, understanding is achieved when readers are able to activate and apply their text-appropriate world knowledge in order to generate inferences regarding what the text is about. Furthermore, readers use their world knowledge in order to infer how elements of a text are related, and thus ensure that their mental representation of the text is coherent. Although there is consensus among discourse-processing researchers regarding the importance of world knowledge in understanding and achieving text coherence, there has been considerable theoretical conjecture and empirical research regarding what kinds of knowledge-based inferences readers normally generate during comprehension (Graesser, Singer, & Trabasso, 1994; Magliano & Graesser, 1991; McKoon & Ratcliff, 1992). Much of this debate has centered around (a) what inferences are necessary for achieving text coherence, (b) the extent to which readers generate inferences that are not necessary for achieving coherence, and more recently (c) what processes are involved in making inferences available to a reader. Although there has been much progress, the debates regarding these issues are far from resolved.
