Abstract
An interactive conception of the psychological refractory period (PRP) effect is proposed on the basis of Hommel's (1998) two-process approach to account for compatibility effects in the PRP task. The interactive conception account assumes that response selection has two components. One component is stimulus-response (S-R) translation, which can occur automatically and simultaneously for both tasks. The other component is final response selection, which is the locus of the bottleneck and can process only one task at a time. The account suggests that between-task crosstalk and noncurrent-task response association have strong impacts on S-R translation when there is a contingency between the two tasks. Six PRP experiments were conducted: The first three experiments contained no contingency between the two tasks, but the last three did. Greenwald and Shulman's (1973) S-R compatible and ideomotor compatible tasks were used in Experiments 1-3, with both responses (R1 and R2 for Task 1 and Task 2, respectively) being required in Experiments 1-2 and only R2 in Experiment 3. The interactive conception predicts that the PRP effect should be obtained in Experiments 1 and 2 but not in Experiment 3 because the selection of R2 has to wait until the selection of R1 is completed. A PRP effect was evident in Experiments 1 and 2 but not Experiment 3. In Experiment 4, the dimensional overlap of the color between the first stimulus (S1) and the second one (S2) was manipulated and participants were instructed to respond to S2 only. A large PRP effect was obtained for the dimensional overlap condition and a small, but significant, PRP effect for the condition with no dimensional overlap. Experiments 5 and 6 examined the effect of S1-S2 correlation (high, low, and neutral), as well as spatial correspondence (R1-R2 correspondent and R1-R2 noncorrespondent) in Experiment 6, on the PRP effect. An overaddictive interaction of S1-S2 correlation and SOA was obtained for both experiments. A comparison between Experiment 5 and 6 showed no difference in the PRP effect obtained with the three levels of S1-S2 correlation. However, the effect of correlation tended to be larger at the short SOA in Experiment 6, in which spatial correspondence of responses was manipulated, than in Experiment 5, in which it was not. Results of these experiments are in agreement with the interactive conception of the PRP effect, in which the contingency between two tasks affects the S-R translation processing, which is distinct from the processing of final response selection.
