Piaget proposed that understanding permanency, understanding occlusion events, and forming mental representations were synonymous; however, accumulating evidence indicates that those concepts are "not" unified in development. Infants reach for endarkened objects at younger ages than for occluded objects, and infants' looking patterns suggest that they expect occluded objects to reappear at younger ages than they reach for them. We reaffirm the latter finding in 5- to 6-month-olds and find similar responses to faded objects, but we fail to find that pattern in response to endarkened objects. This suggests that looking behavior and reaching behavior are both sensitive to method of disappearance, but with opposite effects. Current cognition-oriented (i.e. representation-oriented) explanations of looking behavior cannot easily accommodate these results; neither can perceptual-preference explanations, nor the traditional ecological reinterpretations of object permanence. A revised ecological hypothesis, invoking affordance learning, suggests how these differences could arise developmentally.