It is clear that the outcome for schizophrenia is better in developing than in industrialized countries (see Hopper this volume) yet attempts to account for this difference have been speculative and elaborated in “a virtual ethnographic vacuum” (Hopper 1992:95). It is also clear that family predictors of relapse – hostility, criticism, and emotional overinvolvement expressed by relatives toward the ill family member – have proved robust when tested cross-culturally (Jenkins and Karno 1992). Jenkins (1991) has urged attention to the cultural salience and meaning of these predictors of relapse. Ethnographic study of particular families in the developing world coping with psychotic illness in the household is recommended to explain how social and emotional factors might moderate prognosis (Corin 1990; Sartorius 1992; Lucas and Barrett 1995).

Hostility in familial interactions is identified when criticism is generalized or pervasive (for example, “he's a failure at everything he does”) or rejection (“he can live on the streets for all I care”). Criticism is defined in terms of a negative affective response (usually anger) to rule violation observed in language content and paralinguistic features of speech (Vaughn and Leff 1976). It is necessary both to identify cultural rules and to appreciate the range of familial and community responses to rule violation in order to understand criticism and hostility. Emotional overinvolvement must be understood in terms of kin relations and notions of the self.