(1987) see theories of diffusion and of economic demand as mutually exclusive, while Montgomery and Casterline (1996) see them as empirically indistinguishable and mutually reinforcing, but the two approaches always are opposed conceptually.¹ Microeconomic theories attend to the choices of representative individuals or couples abstracted from their social settings. Diffusion theories attend to communities in which people interact. Observing, discussing, criticizing, and evaluating, people pass information from one to another and from public sources to groups. Communication along interpersonal channels and through impersonal media provides information about “the existence of new behavioral options,” narrows “the range of uncertainty regarding the consequences of new choices,” and “reduce[s] the costs of innovation” by modifying social norms (Montgomery and Chung, 1999:181). Learning is social as well as individual.

Attention to the role of social processes in fertility change has led to renewed interest in community-level effects.² In microeconomic accounts of fertility change, structural characteristics of communities such as the level of nonagricultural employment, literacy, and accessibility are conceived of as determining the costs and benefits of children and the costs of fertility regulation. In accounts of fertility change based on diffusion, social learning may produce a process of endogenous feedback that causes changes in fertility to outpace changes in socioeconomic determinants. Influenced by one another, the members of a community also may develop distinctive patterns of contraceptive use. The boundaries between communities, whether ethnic, linguistic, or cultural, are seen as impeding the flow of communication, thus setting communities on divergent paths of fertility change.

All of this points us toward a socially informed theory of fertility change, but it remains dependent on outdated concepts of culture. Paralleling Hammers (1990:456) “agenda…for a culturally smart microeconomics,”³ this paper sketches an approach to research on diffusion informed by contemporary developments in the theory of culture.⁴ The first section of the paper briefly reviews the place of diffusion in three moments of twentieth-century anthropology: early studies of the history and geographical distribution of cultural traits, mid-century studies of the structure and function of sociocultural systems, and the more recent turn from structural functionalism to practice. The second section argues that key anthropological studies, largely in the classic structural functional mold, undercut or sharply qualify two key assumptions of theories of diffusion based on imitation or contagion: that knowledge and ideas concerning birth control are likely to be novel and that they remain unchanged as they spread from one culture to another. The third and fourth sections, based on contemporary developments in practice theory, outline