The overwhelming majority of research on environmental influences has concentrated on the social environment, to the relative neglect of the physical environment. This neglect is justified by an unvalidated hypothesis, namely that the physical environment must be mediated by social parameters in order to influence development. Two studies were conducted to test that hypothesis. Subjects in both studies were 12-month-old infants. Physical and social environment codes, based on direct home observation, were taken from the Purdue Home Stimulation Inventory. The outcome variable of the first study was that of language development, while that of the second was infant mastery motivation. In both studies, hierarchical regression was used to assess whether unique predictive variance was associated with either the physical or social environments. Results indicated that when social environment codes were entered first, the physical environment codes still maintained their predictive power. In contrast, when physical environment codes were entered first, the predictive power of the social environment codes dropped to nonsignificance. These results are contrary to the hypothesis that the physical environment must be mediated by the social environment, and suggest that neglecting the study of physical environment influences cannot be justified empirically. (RH)