In an extension of Gibson’s (1979) concept of object affordance, it has been proposed that motor representations are automatically evoked by pictures of graspable objects. A variety of effects on left/right-handed keypress responses to the perceptual attributes of such images have been taken as evidence that features of actions, including the hand best suited to grasp an object, contribute to the effect of the handle’s left/right location on response selection. We present an argument against this claim by establishing that all of these effects are based on spatial codes, including effects mistakenly interpreted to reflect the influence of limb-specific features of a grasp action. We also present 6 experiments showing that under certain task conditions, limb-specific effects on response selection are indeed automatically generated by the task-irrelevant image of a graspable object. These effects are found either when the observer makes keypress or reach-and-grasp responses to the laterality of a pictured hand superimposed on a depicted object. Both tasks recruit control processes that determine how the hand is selected and configured to grasp an object. We infer that processes implicated in the planning of a reach-and-grasp action themselves determine whether the task-irrelevant picture of an object triggers motor-based rather than spatial features. Our results have deep implications for the widely used concept of an affordance for action furnished by an object.