Is the observation of the human kinematics sufficient to activate automatic imitation of transitive actions?