My purpose is to evaluate the naturalistic theory of content proposed by Jerry Fodor. This is an informational theory which attempts to solve the disjunction problem that notoriously plagues such theories by appealing to the relationship of asymmetric dependence. I argue that this theory fails to apply to two important categories of mental state. On the one hand, it fails to apply to such personal-level states as beliefs, since the required laws do not hold. This is because of the cognitive penetrability of such processes as object recognition. On the other hand, it fails to apply to subpersonal representational states of the kind that figure in cognitive psychological explanations of our cognitive and perceptual capacities. Here the required laws hold but there is no relationship of asymmetric dependency between the various laws that govern the tokening of any particular type of sub-personal state.