ABSTRACT
Many studies have shown that remembering and learning from text depend on both textual characteristics and the cognitive properties of readers (Denhière & Rossi, 1991). Kintsch (1994) demonstrated that with explanatory texts, remembering, and even learning, was better or occurred more quickly when the exposed concepts were just beyond the current state of the reader’s knowledge. McNamara, Kintsch, Songer, and Kintsch (1996) showed that when the coherence of text was weakened, that is, when nouns were replaced by pronouns, and descriptive elaborations and connectives were removed, the readers with prior knowledge used compensatory comprehension processes to infer the relations not stated in the text. However, according to these authors, readers without relevant prior knowledge need to read a fully coherent, very explicit text in order to construct an efficient representation of the text base, which may in turn be necessary for the subsequent construction of a situation model (van Dijk & Kintsch, 1983). Kintsch and Franzke (1995) studied the construction of a situation model about the civil war in Sri Lanka by subjects with various degrees of special knowledge (i.e., the political goal of the war) and with the same level of general knowledge (i.e., the war’s schema). They found that only subjects who were given information about the political goal of the war were able to reproduce a fair amount of information about the political situation and to form an adequate situation model. They concluded that “the reader’s war schema was sufficient to understand one part of the text, at least superficially, whereas the politics schema was useless by itself and needed to be combined with specific prior information in order to support comprehension” (p. 331).
