Abstract
Hybrid distractor-interference tasks frequently produce a paradoxical data pattern. Within-trial congruency effects associated with two distractors interact, consistent with domain-general control, while across-trial congruency sequence effects (CSEs) associated with the same distractors sum additively, consistent with domain-specific control. Recent findings from the hybrid prime-Simon task suggest that controlling for a confound called the reactivation aversion effect (RAE) resolves this "generality-specificity paradox" by eliminating the within-trial interaction between congruency effects. As there are competing accounts of this interaction, however, it remains unclear whether controlling for the RAE always eliminates it. To contrast this possibility with the competing accounts, we varied across three experiments (N = 168) (a) whether one of the two congruency effects was relatively large or relatively small and (b) whether two distractors engendered the same or different types of conflict. We observed a within-trial interaction while controlling for the RAE when one of the two congruency effects was relatively large regardless of whether the two distractors engendered the same or different types of conflict. This outcome suggests that controlling for the RAE may not always resolve the generality-specificity paradox. It also supports the view that observing a within-trial interaction is more likely when there is sufficient "operating space" for congruency effects to interact.