Temporal locus of inference in the comprehension of brief passages: recognizing and verifying implications about instruments.

Recent investigations of language comprehension have shown that people frequently report recognizing test sentences which express ideas that were only implied in an antecedent passage. The present study asked whether the inferences underlying these false alarms are drawn when the original message is encoded or at test time. In Exp. 1, it was shown that the recognition rate for direct quotations from brief passages is clearly higher than for sentences expressing implications about the instruments used to perform actions. In Exps. 2, 3, and 4 the subjects were timed while they decided whether test sentences (a) had been heard verbatim in the passage or (b) were true concerning the passage. In all conditions, subjects needed about .2 sec. more to respond affirmatively to "implicit" sentences than to ones that quoted or paraphrased the passage. It was concluded that at least a subset of the cognitive processes associated with the inferences in question is executed at test time.