Speech recoding in silent reading

Theoretical controversy surrounds the issue of whether or not silent reading involves speech recoding. This was investigated in two experiments by assessing performance on a Stroop color-word task carried out with subjects either silent or articulating irrelevantly (saying "bla" continuously). It was found that the usual decrement in performance resulting from lack of congruency between ink color and color word was attenuated in the articulation condition. The results provide evidence for the presence of speech recoding in silent reading. As a second test of this hypothesis, the Stroop task was also carried out in conjunction with either a graphemic or a phonemic task. The usual decrement in performance was attenuated more by the phonemic than by the graphemic task, therefore providing further support for the hypothesis. The relationship between individual differences in Stroop performance and those in reading speed and in personality (as assessed by the Eysenck Personality Questionnaire) were also examined. © 1978 Psychonomic Society, Inc.