Spontaneous Speech Generation

Abstract

A speech generation task was performed by bilingual subjects whilst they received irrelevant messages in one ear. The irrelevant messages varied in content as well as in language. Although these messages appeared to have a significant influence on the rate of speech, this variation was not consistent with any of the three hypotheses suggested. A further speech generation experiment was then carried out in which subjects received as irrelevant auditory input, in one ear, either a prose passage or “emotional” words repeated over a period of 1 min. Although the rate of speech did not seem to be affected by the irrelevant input, a memory test for words spoken, as distinct from words heard seemed to indicate that the “emotional” words were significantly better recalled than the words from the prose passage. When a control experiment was performed with prose or repeated neutral words as auditory input, no such difference in recall was obtained. This result was seen as favouring Deutsch's model of the blocking of irrelevant speech.