| Three experiments investigated whether the Braine, Reiser, and Rumain (1984) model of natural propositional reasoning describes the reasoning of adults and children, and whether the inferences of the model operate in text comprehension and comprehension monitoring. Experiment 1 tested the model in the context of an abstract reasoning task by presenting 16 adults, 16 seventh graders, and 16 fourth graders with premisses designed to elicit 1- and 2-step chains of inferences; subjects were asked to write down everything they could infer from the premisses and/or to evaluate given conclusions. Experiments 2 and 3 assessed the model in the context of meaningful text. In Experiment 2, 54 adults, 54 seventh graders, and 54 fourth graders read a set of stories designed to elicit inferences predicted by the model. They were asked to evaluate the conclusion of each story (a task that required either zero, one, or two model-predicted inferences); they then had to decide whether several pieces of information (actual paraphrases, model-predicted inferences, and other, non-predicted logical inferences) were paraphrases of information explicitly given in the story or had to be inferred. In Experiment 3, 52 adults and 33 fourth graders were shown two story fragments from Experiment 2 and asked to write endings to the stories. Together, the results of the three experiments supported the hypotheses that children as young as fourth graders use reasoning of the sort described by the model (with the exception of one inference schema) and that these inferences are at work in both text comprehension and comprehension monitoring. |
