Prevendo Inferências Lógico-Predicativas no Processamento de Texto

Abstract

Previous investigations had established that sentential-logic inferences are made effortlessly during text processing. Three experiments extended this investigation to predicate-logic inferences in text comprehension. Experiments 1 and 2 presented stories in which premises for a single predicate-logic inference were embedded; Experiment 3 presented stories containing premises for multiple predicate-logic inferences. Participants made almost no errors in judging whether a final sentence to each story was sensible, showing that the predicate-logic inferences were made during reading. Further, participants judged that these inferences as being similar to paraphrases, showing that they did not think that any inference had been made. Thus, although predicate-logic requires complex representations of the internal structure of propositions that is lacking in sentential logic, the present results show that their application during reading is equally effortless. The theoretical implications of the results both for the text comprehension literature and for the logical reasoning literature are discussed.