Three experiments investigating the effects of response production and inhibition on the N2 and P3 components of the ERP are reported. In the first experiment, 12 young female volunteers were presented with the words "push' and "wait' (semantic stimuli). On a separate series of trials, they were presented with arbitrary symbols assigned the same meanings (symbolic stimuli). For each stimulus series half of the stimuli were degraded. To obtain an estimate of reliability of the data, each task was repeated. Data were collected from Fz, Cz and Pz electrode sites. The P3 amplitude had a parietal maximum when the stimuli instructed subjects to respond (Go). The P3 was equal at central and parietal sites when the stimuli instructed the subjects to withhold a response (No-Go). This topographic pattern was obtained for all stimulus manipulations, simple and degraded stimuli, words and symbols, and for the first and second runs. The N2 was a frontal maximum component that was larger to the No-Go than to the Go stimuli. This result was also robust to the manipulations. A second experiment investigated the dependency of these findings on an overt motor response. In this experiment, the symbolic and semantic stimulus series were each presented twice. The subjects counted the Go stimuli and did not count the No-Go stimuli for one presentation and pressed the reaction time button as in experiment 1 for the other presentation. While counting (compared to button pressing) delayed the N2 and P3 peaks, counting and pressing produced similar results, including the Go/No-Go P3 distribution effects. A third experiment investigated the sensitivity of these findings to the orientation of the symbols instructing the subjects to respond or withhold the response. Again the pattern of results was robust to this manipulation.