Integration of Semantic and Structural Constraints in Narrative Comprehension

The original debate between semantic and syntactic approaches to issues in reading comprehension has lost some of its appeal as psycholinguistics has embraced various interactive models of reading comprehension. This study, based on Information Integration Theory (Anderson, 1981, 1982) and its methodological counterpart, Functional Measurement, investigates whether semantic and structural information may be integrated during the comprehension process by means of a simple algebraic rule and whether the comprehension task may introduce a systematic bias to this process. Twelve high school students participated in a series of experiments which attempted to cross-validate the basic findings. Subjects were asked metalinguistic questions dealing with the logic and organizational nature of the story to be comprehended. The logic of the story was manipulated by varying the relative compatibility of an industrious protagonist and his report of having done ‘all/some/none’ of the work assigned to him. The structural manipulation involved presenting the sentences of the story in a fully mixed, partially mixed or nonmixed order, according to a story grammar model. These manipulations were performed within the methodological framework of a complete two-factorial design recommended by Functional Measurement. Findings indicate that the semantic and structural pieces of information were integrated by an equal weighted averaging rule, this process being modified by the comprehension task.