A subject with complete and chronic absence of cutaneous and proprioceptive feedback applied inefficiently elevated grip forces to lift and hold instrumented objects of varying weight. However, the ability to differentially scale grip forces according to various weights (0.36 kg, 0.61 kg and 0.86 kg) was mainly maintained when the subject was allowed to lift a constant weight repeatedly under constant visual control. The normally very precise temporal coupling between grip and load forces was substantially disturbed for lifts performed by the deafferented subject. In particular, maximum grip force lagging some 100 ms behind maximum load, suggesting rather feedback-based than predictive grip force control. Our findings demonstrate the importance of cutaneous and proprioceptive afferents for predictive regulation of prehensile finger forces. However, alternative sensory signals, e.g. from visual cues, small myelinated and unmyelinated muscle afferents or neck proprioception, may provide some provisional grip force control in the absence of cutaneous and proprioceptive feedback.