The effects of concurrent tasks on reading: Implications for phonological recoding

Abstract

The use of phonological codes in normal reading was investigated in three experiments in which subjects were asked to read and comprehend prose passages while simultaneously performing concurrent tasks. These tasks differed in terms of stimulus type (verbal-nonverbal), response mode (articulatory-nonarticulatory), and stimulus-response congruence (transformation required-not required). A statistical adjustment procedure was used to take into account the general processing capacity required by the concurrent tasks. After the general capacity demands of these tasks were measured and partialled out, their specific interference with reading was shown to be related to whether or not the input required transformation, and to be unrelated to stimulus type or response mode. Concurrent tasks, such as shadowing, which required an articulatory response but no transformation, only interfered with normal reading when the processing demands of the concurrent task were not taken into account. This suggests that previous findings in which shadowing interfered with normal reading may have been due to a reduction in general processing capacity by the concurrent task, rather than to speech-specific interference. A fourth experiment showed that shadowing did interfere when the reading task was structured so as to force subjects to use phonological codes. Since shadowing did have the potential to interfere with the reader's use of phonological codes, but did not interfere with normal reading, the results suggest that skilled reading may not require the use of phonological codes.