Evidence for a Time Constant in Rate of Speech under Delayed Auditory Feedback

Shadowing at one and at two ears

The assumption that speech shadowing commandeers all the subject's capacity was tested in two ways: by manipulating the linguistic parameters of the non-attended message, and by superimposing a task of repeating the second message whenever possible while shadowing the first message. Eight subjects shadowed prose or random words (130 words per min). Non-attended messages of either type were also presented as forward, or reversed speech. With reversed speech, percentage shadowing scores for random words were less than prose (P < 0.001). Random word shadowing was further adversely affected with non-attended, forward speech, there being a significant interaction (P < 0.001) between shadowed message type and intelligibility of the non-attended message. Subjects repeated more words shadowing both messages than one alone (P < 0.001), despite the switching which this entailed, and shadowing scores were related to the redundancy of both message combinations.