The detection of a conflict between relevant and irrelevant information on a given trial typically results in a smaller conflict effect on the next trial. This sequential effect has been interpreted as an expression of cognitive control implemented to resolve conflict. In this context, 2 different but related issues have received increasing attention in the literature. The first issue is whether the detection of motor conflict is necessary to induce cognitive control or, alternatively, whether the detection of perceptual conflict is sufficient. The second issue concerns whether awareness of the conflict is necessary to induce cognitive control. Here, we address both issues in a single design. Our reaction-time (RT) results indicate that conflict-driven control is domain-specific. The detection of perceptual conflict on the previous trial selectively reduces perceptual conflict on the next trial. Similarly, the detection of motor conflict on the previous trial selectively reduces motor conflict on the next trial. For errors, adaptive control seemed to be more general: The detection of perceptual or motor conflict on the previous trial reduced the frequency of errors on response-conflict trials. Furthermore, unconsciously triggered conflict adaptation was observed, but not systematically. Results on errors provide some evidence that sensitivity to an unconscious conflict on the previous trial reduces the frequency of errors on the current trial. For RT analyses however, unconscious conflict appeared not to be sufficient to induce cognitive control. This pattern of results is in line with previous studies examining the role of consciousness in conflict adaptation.