This study tested the prediction that covert speech behavior measured electromyographically from the lips is significantly more prominent during a brief silent-language recitation task than a brief nonlanguage visualization task. Subjects were 20 right-handed, adult volunteers who agreed to participate. Subjects were tested in a multiple-baseline reversal design following an ABAB procedure whereby A1 and A2 were 30-sec. rest periods. B1 and B2 were alternatively assigned 30-sec. silent-language recitation and visualization test periods, respectively. Subjects' dorsal lips and nondominant forearm EMG measures were taken during resting baseline and testing conditions. In addition, subjects' skin surface temperature and heartrate were measured during the rest and test conditions. For the silent-language task, subjects were asked to recite 'mentally' the Pledge of Allegiance to the flag. Subjects were instructed to 'imagine seeing' the American flag for the visualization task. Subjects' mean lip EMG activity increased significantly from rest to the silent-language recitation task, while no significant change in mean lip EMG was observed from rest to the visualization condition.