Abstract
This paper outlines three problems that are foreseen to face us as we enter the 1990s. These are: (1) defining the nature of relationships in the face of discrepancies between the reports of partners and outsiders about the events that occur in the relationship; (2) clarifying the nature of claims that relationships are `processes'; (3) providing a better understanding of the everyday conduct and routines of relationships, particularly the role of everyday talk. It is argued that longitudinal work and work capturing the experiences and reports of both members of a dyad must solve the first problem at a theoretical level before they can be helpful. It is suggested that, after Billig's (1987) argument that thinking and arguing characterize social behavior, relationships are best conceived phenomenally as unfinished business, just as research usually is. Several suggestions are made about the nature of relationships and, in the course of the paper, some suggestions are made for future work on social and personal relationships, loneliness, social support, children's friendship and process models of relating.

This publication has 16 references indexed in Scilit: