The ability to reach and "grasp" (grip or touch) structures for support in reaction to instability is an important element of the postural repertoire. It is unclear, however, how the central nervous system (CNS) resolves the potential conflict between holding an object and the need to release the held object and grasp alternative support, particularly if the held object is perceived to be relevant to the task of stabilizing the body, e.g. an assistive device. This study examined whether compensatory grasping is inhibited when holding an object, and whether the influence differs when holding an assistive device (cane) versus a task-irrelevant object (top handle portion of a cane). We also investigated the influence of preloading the assistive device, to determine whether conflicting demands for arm-muscle activation (requiring disengagement of ongoing agonist or antagonist activity) would influence the inhibition of compensatory grasping. Unpredictable forward and backward platform translations were used to evoke the balancing reactions in 16 healthy young adults. A handrail was mounted to the right and foot motion was constrained by barriers, with the intent that successful balance recovery would (in large-perturbation trials) require subjects to release the held object and contact the rail with the right hand. Results showed that grasping reactions were commonly used to recover equilibrium when the hand was free (rail contact in 71% of large-perturbation trials). However, holding either the cane or canetop had a potent modulating effect: although early biceps activation was almost never inhibited completely (significant activity within 200 ms in 98% of trials), the average activation amplitude was attenuated by 30-64% and the average frequency of handrail contact was reduced by a factor of two or more. This reduced use of the rail occurred even though the consequence often involved falling against a safety harness or barriers. Handrail contact occurred least frequently when holding the cane during forward loss of balance: subjects persisted in pushing on the cane (failing to use the rail) in 93% of trials, even when the perturbations were too large to allow this strategy to be successful. Prior contraction (preloading the cane) did not influence any of these findings. Complex strategies (e.g. partial release of object) were often adopted to allow balance to be recovered without dropping the held object. Remarkably, it appears that the CNS may give priority to the ongoing task of holding an object, even when it has no stabilizing value (cane during backward falls) or any intrinsic value whatsoever (canetop).