Human social systems are constituted at different scales; the local community, while an important organizational unit cross-culturally, has received limited archaeological study. We argue that community-centric studies in archaeology have significant potential in documenting the diversity of small-scale agricultural systems and in promoting comparative study of local societies. Using comparative data from social anthropology, we put forth a definition of community useful for the archaeologist based on three irreducible elements: social reproduction, agricultural production, and self-identification. These elements provide an ideal framework for cross-cultural comparison. We also argue that using a community approach requires certain methodological refinements, such as adopting an appropriate research scale, conducting intensive surface survey, and using analytical strategies such as labor investment and boundary maintenance. We present recently collected data from two prehistoric agricultural communities, Waiohuli in Hawai'i and Tsikwaiye in northern New Mexico, in order to illustrate the strength of this mode of research.