Conventional approaches to family studies typically begin with an understanding or definition of the family that specifies its characteristics as a particular kind of group. This seems eminently reasonable, both commonsensically and as a social scientific practice. We all believe families to inhabit everyday life as concrete entities, and to study them we must clearly designate what is being observed. But how is the student of the family to take his or her encounters with people in the "real world" whose images of family seem radically different from the academic definitions? Consider, for example, the now-familiar refrain of the athlete who explains his team's success by noting that "We have a good family atmosphere going" (Gubrium & Holstein, 1990, p. 140). Or what do we make of urban anthropologist Carol Stack's (1974, p. 58) report that in the community she studied, "When friends more than adequately share the exchange of goods and services, they are called kinsmen … if two women of the same age are helping one another, they call their friend "just a sister""?