It has long been observed that patients with Parkinson's disease (PD) can sometimes react and move quickly in response to an external stimulus in a way that they cannot when required to initiate the movement themselves. This curious phenomenon has sometimes been called 'paradoxical kinesis'. The present study was an attempt to demonstrate this phenomenon in patients with PD using an objective and quantifiable experimental procedure. A reaction time task was used in which participants had to press one of two computer keys, either left or right, to save a cartoon person on a computer screen from being run over by a motor car. In one condition, trials started after a traffic light appeared on the computer screen and then changed from red to green. In a second condition, the participants had to first press a third response key which resulted in the traffic light appearing on screen and changing from red to green. Participants also received both these conditions with the addition of a visual cue, an arrow, which told them in advance which direction to respond in (i.e., left or right key) on each trial. The purpose of the visual cue was to separate the effects of motor planning from motor activation. Healthy controls reacted quickest when they initiated trials themselves whereas the PD group were quicker to respond when trials were externally generated. Both groups were quicker under the visual cue condition. The results are discussed in terms of recent research which has suggested that two separate neural systems may be involved in externally generated or stereotyped actions and motor responses which require self-generated or nonroutine decision making.