The fate of originally presented surface information following recall errors in sentence memory tasks

In three experiments, a word-fragment completion task was used to test three hypotheses about the accessibility of surface information in sentences. Subjects read sentences which were known (Brewer, 1975; 1977) to yield lexical substitutions during a cued-recall task. Following recall, subjects completed word fragments of originally studied words. No reliable facilitation in fragment completion occurred for original words that had been misrecalled, whereas facilitation did occur for original words that were not misrecalled. Moreover, originally studied words that were misrecalled showed fragment completion rates that were similar to items that had never been seen before. These results are consistent with the hypothesis that the original surface information was made inaccessible by the alternative information generated during recall.