Systematic error in the organization of physical action

Current views of the control of complex, purposeful movements acknowledge t hat organizational processes must reconcile multiple concerns. The central priority is of course accomplishing the actor's goal. But in specifying the manner in which this occurs, the action plan must accommodate such factors as the interaction of mechanical forces associated with the motion of a mu ltilinked system (classical mechanics) and, in many cases, intrinsic bias t oward preferred movement patterns, characterized by so-called "coordination dynamics." The most familiar example of the latter is the symmetry constra int, where spatial trajectories and/or temporal landmarks (e.g., reversal p oints) of concurrently-moving body segments (limbs, digits, etc.) exhibit m utual attraction. The natural coordination tendencies that emerge through t hese constraints can facilitate or hinder motor control, depending on the d egree of congruency with the desired movement pattern. Motor control theori sts have long recognized the role of classical mechanics in theories of mov ement organization, but an appreciation of the importance of intrinsic inte rlimb bias has been gained only recently. Although detailed descriptions of temporal coordination dynamics have been provided, systematic attempts to identify additional salient dimensions of interlimb constraint have been lacking. We develop and implement here a nov el method for examining this problem by exploiting two robust principles of psychomotor behavior, the symmetry constraint and the Two-Thirds Power Law . Empirical evidence is provided that the relative spatial patterns of conc urrently moving limbs are naturally constrained in much the same manner as previously identified temporal constraints and, further, that apparent velo city interference is an indirect, secondary consequence of primary spatial assimilation. The theoretical implications of spatial interference are elab orated with respect to movement organization and motor learning. The need t o carefully consider the appropriate dimensions with which to characterize coordination dynamics is also discussed. (C) 2001 Cognitive Science Society , Inc. All rights reserved.