Several studies examining spatial attention have found a discrepancy regarding the effects of exogenous cues on reaction times in visual detection and discrimination tasks. Namely, across a wide range of cue-target intervals, responses are slower for targets at cued than at uncued locations (inhibition of return) in detection tasks, whereas responses are faster for targets at cued than at uncued locations (facilitation) in discrimination tasks. Two hypotheses were proposed to account for this discrepancy. First, attention may dwell much longer on the exogenously cued location in discrimination tasks because stimuli have to be identified (i.e., the delayed attention withdrawal hypothesis). Secondly, due to increased motor preparation in detection tasks, cue-induced motor inhibition may rise much faster in these tasks than in discrimination tasks (i.e., the speeded motor inhibition hypothesis). We examined to what extent these hypotheses can account for effects of exogenous cues in a detection and discrimination task on the extrastriate P1 component, and the onset of motor activation, as indexed by the lateralized readiness potential. Some support was found for the delayed attention withdrawal hypothesis, as task-dependent cueing effects were found on the P1 component. Other aspects of our data, however, indicate that motor inhibition is also involved. Based on these findings, we propose that effects of exogenous cues in detection and discrimination tasks are determined by the interplay between two mechanisms, of which the time courses of activation may be modulated by the specific setting.