The influence of social relationships on human development and behavior is receiving increased attention from psychologists, who are central contributors to the rapidly developing multidisciplinary field of relationship science. In this article, the authors selectively review some of the significant strides that have been made toward understanding the effects of relationships on development and behavior and the processes by which relationships exert their influence on these, with the purpose of highlighting important questions that remain to be answered, controversial issues that need to be resolved, and potentially profitable paths for future inquiry. The authors' thesis is that important advances in psychological knowledge will be achieved from concerted investigation of the relationship context in which most important human behaviors are developed and displayed.

Because interpersonal relationships are the foundation and theme of human life, most human behavior takes place in the context of the individual's relationships with others. Psychologists actively participating in the multidisciplinary effort to develop a science of relationships (Berscheid, 1999) are doing so because they believe that the human's omnipresent relationship context strongly influences each individual's behavior and his or her development over the life span. As a consequence, they believe that a science of human behavior and development that neglects the influence of the individual's interpersonal relationships is destined to be inaccurate and incomplete (see Kelley, 1983).

Types of Relationship

The concept of relationship, even the concept of close relationship, encompasses many different nominal types of relationships (e.g., romantic, parental, friendship, coworker, neighbor). An important question facing relationship science is whether the similarities underlying different types of relationships are sufficient to develop a superordinate body of relationship knowledge or whether relatively independent bodies of knowledge, each addressed to a specific nominal relationship type, are required (Berscheid, 1994; M. S. Clark & Reis, 1988). The answer to this question hinges on two considerations: first, whether mechanisms that serve to organize different categories of relationships (and not just close relationships) can be identified, and second, the extent to which the laws governing behavior with the same partner differ depending on the type of relationship in which interaction takes place (e.g., in a romantic or employment-supervisory context). Such similarities and differences in behavior due to relationship type have yet to be systematically catalogued. Indeed, attempts to develop taxonomies of relationship type are relatively recent, although some progress has been made in identifying commonalities underlying certain subsets of relationships and the functional properties that differentiate them. Many taxonomic models incorporate the early work of M. S. Clark and Mills (1979, in press) who have gathered a great deal of evidence to support their distinction between communal relationships, in which people respond to the other's needs, and exchange relationships, in which benefits are exchanged in repayment for prior benefits or in expectation of future benefits. In one such model, which adopts the notion of modularity from cognitive psychology (Hirschfeld & Gelman, 1994), Bugental and Goodnow (1998) argued that over evolutionary time, certain relationship "domains" became incorporated as discrete modules in human neural architecture. (See de Waai, 1996, for a similar argument conceming relationship structures among primates.) According to Bugental and Goodnow (1998), these relationship domains represent distinct "bodies of knowledge that act as guides to partitioning the world and that facilitate the solving of recurring problems faced by organisms within that world" (p. 400). In contrast to the traditional view that socialization consists of the individual learning one set of principles that is then applied to all social situations, Bugental and Goodnow maintained that socialization is the process of learning the "distinctive sensitivities and regulatory processes" (p. 400) appropriate to different social domains (to which we would add the task of discerning which relationships should be parsed into which domain)

Bugental (2000) has proposed five social domains: (a) an attachment domain, characterized by proximity-maintenance within a protective relationship; (b) a hierarchical power domain, characterized by use and recognition of social dominance; (c) a coalitional group domain, which concerns the identification and maintenance of lines that divide "us" and "them"; (d) a reciprocity domain, characterized by the negotiation of matched benefits with functional equals; and (e) a mating domain, concerned with the selection and protection of access to sexual partners. Each domain is theorized to be distinguished not only by its distinctive cognitive representations but also by components that regulate emotion and social behavior. Thus, each domain is proposed to differ functionally from the others by differing sensitivities to certain social cues and by different operating principles. An earlier and somewhat similar model was proposed by A. P. Fiske (1992), who also argued that relationships can be differentiated into discrete structures: there are no intermediate forms and they are not reducible to any set of continuous dimensions .... They are fundamental and they are also incommensurable, in the sense that there is no general, systematic, higher-level schema that mediates among them. (A. P. Fiske, Kitayama, Markus, & Nisbett, 1998, pp. 950-951)

Relationship taxonomies are more than a useful organizing descriptive tool; they help highlight important questions about the characteristic rules and processes that govern interaction, and thus, these recent taxonomic efforts represent a valuable advance. Nevertheless, it must be noted that typological approaches are sometimes misleading; under certain circumstances, the underlying phenomena may be better represented by continuous dimensions (N. G. Waller & Meehl, 1998). Fraley and Waller's (1998) finding that attachment styles, long construed as categorical, are more accurately viewed as dimensional is instructive in this regard.

Some dimensions underlying relationships have been tentatively identified. For example, Wish, Deutsch, and Kaplan's (1976) multidimensional scaling studies identified four dimensions that appear to underlie people's characterizations of their relationships: cooperative/friendly versus competitive/hostile, equal versus unequal status, intense versus superficial, and socioemotional/informai versus task-oriented/formal. Dimensional models, although inconsistent with the logic of modularity, do not necessarily argue against the value of examining discrete relationship types. It might be useful, however, to investigate whether such types might be better viewed as exemplifying extremes of dimensions rather than as discrete modules (e.g., romantic relationships occupy the cooperative-equal status-intense-socioemotional poles of Wish et al,'s four dimensions). Finally, it must be noted that none of the relationship taxonomies advanced to date has yet generated sufficient empirical research to allow confidence that the interaction patterns said to uniquely characterize each domain actually do so. Obtaining that evidence is made difficult by the individualistic approach to the study of behavior that dominates psychology.

In the study of relationships, the systems perspective would acknowledge that:

1. From the moment of conception, individuals are nested in social relationships that influence the nature and operation of the many hierarchically organized biological and behavioral systems each individual encompasses. 2. Each relationship is itself nested in a social environmental system and in a physical environmental system, which together represent each relationship's ecological niche. 3. The specific ecological niche of each relationship is, in turn, embedded in larger societal and cultural systems (see, e.g., Levinger, 1994). 4. All of these systems are simultaneously evolving and influencing each other over time. Like biologists, most relationship scientists endorse these four propositions. As a consequence, virtually all believe that relation- ~hip systems are of critical importance to an understanding of human behavior and the course of human development.