When we use our digits to manipulate objects the applied fingertip forces and torques tangential to the grip surfaces are a result of complex muscle activity. These patterns are acquired during our ontogenetic development and we select them according to the manipulative intent. But the basic force coordination expressed in these patterns has to be tuned to the physical properties of the current object, e.g. shape, surface friction and weight. This takes place primarily by parametric adjustments of the force output based on internal models of the target object, i.e. implicit memory systems that represent critical object properties. From visual or haptic information we identify objects and automatically retrieve the relevant models. These models are then used to adapt the motor commands prior to their execution. The formation of models and their swift updating with changes in object properties depend, however, on signals from tactile sensors in the fingertips.