ABSTRACT
Current theories of text comprehension assume three levels of representation of the text and of its content (Fletcher & Chrysler, 1990; Kintsch, 1988; Kintsch & van Dijk, 1978; Tapiero, 1991; van Dijk & Kintsch, 1983). The surface level refers to the representation of the exact wording and syntax a reader captures from a text. The semantic level results from two types of processing: microprocessing and macroprocessing. The microprocessing implies the construction of a locally coherent propositional network called the textbase (see Kintsch, 1974; Kintsch & van Dijk, 1978). The macroprocessing involves propositions of the textbase (micropropositions) organized in a hierarchical and coherent sequence. These two processes, local and global, lead to two levels of discourse organization: the microstructure and the macrostructure. Thus, while reading a text, step by step the reader constructs the microstructure by applying local coherence relations (e.g., referential, and/or temporal and causal relations). The construction of global discourse coherence (macrostructure) is made by restructuring the microstructure into a meaningful global structure. At this level of comprehension, the reader’s goal is to construct a representation that is faithful to the text (Kintsch & van Dijk, 1978; O’Brien & Lorch, 1995; van Dijk & Kintsch, 1983) and there is no need for the intervention of the reader’s prior knowledge.
