Attention to Action: Securing Task-Relevant Control in Spoken Word Production

The Stroop phenomenon is the finding that color naming is inhibited by incongruent color words but word reading not by incongruent colors. When the stimulus onset asynchrony (SOA) is manipulated, maximal inhibition of incongruent words on color naming is obtained when the words are presented within 100 msec of the colors, whereas facilitation of preexposed congruent words is constant. These finding are obtained both with and without task certainty. Whereas existing models explain the basic Stroop effects, they fail to account for the time course findings and for performance under task uncertainty. In this paper, I extend and apply the WEAVER++ model of spoken word production (Roelofs, 1992, 1993, 1997c; Levelt, Roelofs, & Meyer, 1999) to performance on the Stroop task and show that the model accounts for