The present study sought to determine whether the increased postural instability produced by a spoken mental task was due to competing demands for attentional resources or perturbation of posture by articulation. Postural sway was measured in 36 normal subjects under the following conditions: repeating a number aloud (articulation), counting backwards aloud in multiples of seven (articulation and attention), counting backwards silently (attention), and no mental task (neither articulation nor attention). Articulation resulted in a significant increase in sway, whereas no effect of attention was observed. We conclude that in order to accurately assess the effect of attentional demands on postural control, it is important to eliminate or control the effects of articulation.