This study analyzed the ability to coordinate motor behavior in a social context requiring different levels of cooperation. Pairs of subjects seated in front of a wide computer screen and controlled their own (red or green) cursor by an isometric joystick. Each subject had to move its cursor from a central to different peripheral visual targets in a reaction-time paradigm. In the NULL COOPERATION subjects performed the task individually. Under LOW COOPERATION both subjects had to enter the peripheral targets within a given common time window, regardless of the movement trajectory chosen by each of them. In the STRONG COOPERATION the two subjects worked together, and each individual had to scale its action in amplitude and direction to account for the force exerted by the other. Thus, the three conditions were similar in terms of motor behavior, but different in their degree of social cooperation. We tested 13 pairs of healthy volunteers (age: 20–30 years). We found that the reaction-time and the movement-time in the STRONG COOPERATION condition were significantly longer than in both the NULL and LOW condition. The analysis of the cursor's velocity profile also showed a significant decrease in the STRONG condition, as compared to the other two. We conclude that social cooperation through movement requires a rescaling of the basic parameters of each individual motor behavior and, under the condition tested, and a slowdown in time of the performance of each partner.