Comprehension and recall of sentences.

Preliminary experiments showed that 5s better recall a noun pair if they generate their own linking sentence for the pair than if they merely read an equivalent linking sentence. Initial attempts to explain this effect in terms of memory search activities or idiosyncratic ally high-associative mediators proved unproductive in later experiments reported here. hypothesis was then offered that the generate vs. read conditions differ in comprehension of the sentences and that comprehension aids retention. Subsequent experiments on incidental learning showed that recall is excellent when S is set to process a sentence in different ways designed to promote comprehension of its meaning, whereas equivalent exposure to or mouthing of the words in control sentences without comprehension produces relatively little recall. following experiments are concerned with the facilitation of paired-associate learning produced by embedding each word pair in a sentence. Rohwer (1966) found that an 5 who hears a linking sentence such as The COW chased the will recall the COW-BALL pair better than a control 5 who simply studied the pair without a sentence context. In repeating some of Rohwer's paradigms, another phenomenon was uncovered which led into the present experimental series. phenomenon is that 5s better remembered noun pairs embedded in sentences they generated than they did pairs embedded in sentences E gave them. At the time of study or input, 5s in the read condition read aloud a presented sentence (e.g., COW chased the BALL), whereas those in the generation condition saw the pair COW-BALL and had to make up and say aloud a linking sentence. Although input times were controlled, later recall (of BALL when cued with COW) was about 25-30% higher in the generation condition. This result is quite reliable, having been replicated several times in the experiments reported subsequently. Why does generating a linking sentence