Argues that archaeologists should not assume a one-to-one correlation between the number of houses and households. A case study of a contemporary Iranian village is presented where most households do not live in self-contained dwelling units. Methods of recognizing this fragmented household pattern in the archaeological record are discussed. (AM)