The cost of incongruent stimuli is reduced when conflict is expected. This series of experiments tested whether this improved performance is due to repetition priming or to enhanced cognitive control. Using a paradigm in which Word and Number Stroop alternated every trial, Experiment 1 assessed dynamic trial-to-trial changes. Incongruent trials led to task-specific reduction of conflict (trial n ϩ 2) without cross-task modulation (trial n ϩ 1), but this was fully explained by repetition priming. In contrast, an increased ratio of incongruent words did lead to sustained task-specific enhancement, above and beyond repetition priming (Experiment 2). Experiments 3 and 4 assessed the voluntary modulation of cognitive control: A cue predicted the congruency of the incoming trial, allowing participants to establish the correct mindset (Word Stroop in Experiment 3, Flanker task in Experiment 4). Preparing oneself to process an incongruent word (or flanker) enhanced conflict resolution in the subsequent Number Stroop, an example of cross-task modulation. Taken together, these experiments reveal the multifaceted aspects of conflict resolution: Trial-to-trial changes are often due to repetition priming; sustained modulations brought about by task demands are task specific; and voluntary modulations are task general.