Encyclopedia of Sociology Vol

freedom. Given its Meadian theory of self (an active theory of self that chooses between alternatives), humanist sociology is concerned with how this is best realized within a community. Humanist sociology begins with the fundamental assumption that all varieties of humanism hold—that individuals are the measure of all things. Using a nonpositivistic epistemological foundation, humanist sociologist’s employ their methods of research to answer the most important question that can be asked by a humanist sociologist about human behavior, the one originally raised by the Enlightenment philosophes: How can social science help to fashion a humane society in which freedom can best be realized?

———(1781) 1965 The Critique of Pure Reason. New York: St. Martin’s Press.

Lee, Alfred McClung 1978 Sociology for Whom? New York: Oxford University Press.

Locke, John (1690) 1894 Essay Concerning Human Understanding. Oxford: Oxford University Press.

Mead, George Herbert (1934) 1974 Mind, Self and Society. Chicago: University of Chicago Press.

Meltzer, Bernard N., John W. Petras, and Larry T. Reynolds 1977 Symbolic Interactionism: Genesis, Varieties, and Criticisms. Boston: Routledge and Kegan Paul.

Mills, C. Wright 1959 The Sociological Imagination. New York: Oxford University Press.

Randall, John Herman, Jr. 1976 The Making of the Modern Mind. New York: Columbia University Press.

REFERENCES

Dewey, John 1929 The Quest for Certainty. New York: Minton, Balch and Company.

———1931 Context and Thought. Berkeley: University

of California Press.

———1939 Freedom and Culture. New York: G.P.

Putnam’s Sons.

Flynn, Charles 1976 Association for Humanist Sociology Newsletter 1(1) pp 1–2.

Hegel, Georg Wilhelm Freidrich (1807) 1967 The Phenomenology of Mind. New York: Harper Colophon Books.

———(1821) 1967. The Philosophy of Right. Oxford:

Clarendon Press.

Kant, Immanuel (1788) 1949 The Critique of Practical Reason. Chicago: University of Chicago Press.

Rossides, Daniel W. 1998 Social Theory: Its Origins, History, and Contemporary Relevance. Dix Hills, N.J.: General Hall.

Rousseau, Jean Jacques (1755) 1985 Discourse on Human Inequality. New York: Penguin Books.

Scimecca, Joseph A. 1995 Society and Freedom, 2nd ed. Chicago: Nelson-Hall.

Seidman, Steven 1983 Liberalism and the Origins of European Social Thought. Berkeley: University of California Press.

JOSEPH A. SCIMECCA

HYPOTHESIS TESTING

See Scientific Explanation; Statistical Inference.

IDEAL TYPES

See Typologies.

IDENTITY THEORY

Identity theory, in the present context, has its referent in a specific and delimited literature that seeks to develop and empirically examine a theoretical explanation, derived from what has been called a structural symbolic interactionist perspective (Stryker 1980), of role choice behavior. It is only one of a large number of formulations— social scientific, therapeutic, humanistic—in which the concept of identity plays a central role, formulations having their roots in a variety of disciplines ranging from theology through philosophy to political science, psychology, social psychology, and sociology. Further mention of these diverse formulations will be forgone in order to focus on identity theory as specified above; those who desire leads into the literature of sociology and social psychology to which identity theory most closely relates will find them in McCall and Simmons (1978); Stryker (1980); Weigert (1983); Stryker and Statham (1985); Hewitt (1997a, 1997b); MacKinnon (1994); and Burke and Gecas (1995).

The prototypical question addressed by identity theory, phrased illustratively, is: Why is it that one person, given a free weekend afternoon, chooses to take his or her children to the zoo while another person opts to spend that time on the golf course with friends? The language of this prototypical question implies a scope limitation of the

theory that is important to recognize at the outset of the discussion. The theory is intended to apply to situations where alternative courses of action are reasonably, and reasonably equivalently, open to the actor. A defining assumption of the symbolic interactionist theoretical framework is that human beings are actors, not merely reactors. Identity theory shares this assumption, which recognizes the possibility of choice as a ubiquitous feature of human existence. At the same time, however, identity theory recognizes the sociological truth that social structure and social interaction are equally ubiquitous in constraining—not in a strict sense ‘‘determining’’—human action. That constraint is variable. It may be true in an abstract and philosophical sense that people are ‘‘free’’ to act in any way they choose in any situation in which they may find themselves, including choosing to endure great punishment or even death rather than to behave in ways demanded by others; but surely it is entirely reasonable to presume that jailed prisoners have no viable options with respect to many— likely most—facets of life and in any event have fewer viable options than persons who are not jailed. Identity theory has more to say on those the latter persons than on the former, and more to say on those—perhaps few—aspects of life about which the former do have reasonable choice than on those many aspects of prisoner life where options, as a practical matter, do not exist.

As a derivative of a symbolic interactionist theoretical framework, identity theory shares a number of the assumptions or premises of

interactionist thought in general. One, that human beings are actors as well as reactors, has already been suggested. A second is that human action and interaction are critically shaped by definitions or interpretations of the situations of action and interaction, which definitions and interpretations are based on shared meanings developed in the course of interaction with others. A third premise is that the meanings which persons attribute to themselves, their self-conceptions, are especially critical to the process producing their action and interaction. And a fourth premise is that self-conceptions, like other meanings, are shaped in the course of interaction with others and are, at least in the initial instance and at least largely, the outcomes of others’ responses to the person.

The fourth premise has sometimes been phrased as ‘‘self reflects society.’’ Taken in conjunction with the third premise, it gives rise to the basic theoretical proposition or formula of symbolic interactionism: Society shapes self, which shapes social behavior. That formula, it must be noted, admits of and, indeed, insists upon the possibility of reciprocity among its components—social behavior impacts self and society, and self can impact society. Identity theory builds upon refinements of the traditional symbolic interactionist framework and specifications of its basic formula.

The refinements essentially have to do with three facets of the traditional symbolic interactionist framework as it evolved from Mead ([1934] 1962), Cooley ([1902] 1983), Blumer (1969), and others: the conceptualization of society, the conceptualization of self, and the relative weight to be accorded social structure versus interpretive processes in accounts of human behavior. The traditional framework tends to view ‘‘society’’ as unitary, as a relatively undifferentiated and unorganized phenomenon with few, if any, internal barriers to the evolution of universally shared meanings. It also tends to a view ‘‘society’’ as an unstable and ephemeral reflection, even reification, of relatively transient, ever-shifting patterns of interaction. In this view of society, social structures, as these are typically conceived of by sociologists, have little place in accounts of persons’ behaviors. These accounts tend to be innocent of a coherent sense of extant social constraints on those behaviors, and there are few means of linking the dynamics of social

interaction in reasonably precise ways to the broader social settings that serve as context for persons’ action and interaction.

Further, and enlarging this theme of an inadequate conceptualization and consequent neglect of social structure, this view of society tends to dissolve social structure in the universal solvent of subjective definitions and interpretations, thus missing the obdurate reality of social forms whose impact on behavior is undeniable. To say this does not deny the import for social life of the definitional and interpretative processes central to interactionist thinking and explanation. It is, however, to say that seeing these processes as in large degree unanchored and without bounds, as open to any possibility whatsoever without recognizing that some are much more probable than others, results in visualizing social life as less a product of external constraints and more a product of persons’ phenomenology than is likely warranted. Finally, on the premise that self reflects society, this view of society leads directly to a view of self as unitary, as equivalently internally undifferentiated, unorganized, unstable, and ephemeral.

Contemporary sociology’s image of society is considerably different from that contained in traditional symbolic interactionism, and it is the contemporary sociological conceptualization of society that is incorporated into the structural symbolic interactionist frame from which identity theory derives. This contemporary conceptualization emphasizes the durability of the patterned interactions and relationships that are at the heart of sociology’s sense of social structure. It emphasizes social structure’s resistance to change and its tendency to reproduce itself. The contemporary image differs as well by visualizing societies as highly differentiated yet organized systems of interactions and relationships; as complex mosaics of groups, communities, organizations, institutions; and as encompassing a wide variety of crosscutting lines of social demarcation based upon social class, age, gender, ethnicity, religion, and more. This vast diversity of parts is seen as organized in multiple and overlapping ways—interactionally, functionally, and hierarchically. At the same time, the diverse parts of society are taken to be sometimes highly interdependent and sometimes relatively independent of one another, sometimes implicated in close and cooperative interaction and sometimes conflicting.

The symbolic interactionist premise that self reflects society now requires a very different conceptualization of self, one that mirrors the altered conception of society. Self must be seen as multifaceted, as comprised of a variety of parts that are sometimes interdependent and sometimes independent of other parts, sometimes mutually reinforcing and sometimes conflicting, and that are organized in multiple ways. It requires a sense of self in keeping with James’s ([1890] 1990) view that persons have as many selves as there are other persons who react to them, or at least as many as there are groups of others who do so.

Equally important, viewing both society and self as complex and multifaceted as well as organized opens the way to escaping the overly general, almost banal, and essentially untestable qualities of the basic symbolic interactionist formula by permitting theorization of the relations between particular parts of society and particular parts of self, and by permitting reasonable operationalizations of those parts.

In identity theory, this theorization proceeds by specifying the terms of the basic symbolic interactionist formula, doing so by focusing on particulars hypothesized as especially likely to be important in impacting role choice. That is, first of all, the general category of social behavior is specified by taking role choice—opting to pursue action meeting the expectations contained in one role rather than another—as the object of explanation. Role choice is hypothesized to be a consequence of identity salience, a specification of the general category of self, and identity salience is hypothesized to be a consequence of commitment, a specification of society. Identity theory’s fundamental proposition, then, is: Commitment impacts identity salience impacts role choice.

The concept of identity salience develops from the multifaceted view of self articulated above. Self is conceptualized as comprised of a set of discrete identities, or internalized role designations, with persons potentially having as many identities as there are organized systems of role relationships in which they participate. Identities require both that persons be placed as social objects by having others assign a positional designation to them and that the persons accept that designation (Stone 1962; Stryker 1968). By this usage, identities are

self-cognitions tied to roles and, through roles, to positions in organized social relationships; one may speak of the identities of mother, husband, child, doctor, salesman, employee, senator, candidate, priest, tennis player, churchgoer, and so on. By this usage, too, identities are cognitive schemas (Markus 1977), structures of cognitive associations, with the capacity of such schemas to impact ongoing cognitive and perceptual processes (Stryker and Serpe 1994).

Self is not only multifaceted; it is also postulated to be organized. Identity theory takes hierarchy as a principal mode of organization of identities; in particular, it assumes that identities, given their properties as cognitive schemas, will vary in their salience, and that self is a structure of identities organized in a salience hierarchy. Identity salience is defined as the probability that a given identity will be invoked, or called into play, in a variety of situations; alternatively, it can be defined as the differential probability, across persons, that a given identity will be invoked in a given situation. Identity theory’s fundamental proposition hypothesizes that choice between or among behaviors expressive of particular roles will reflect the relative location in the identity salience hierarchy of the identities associated with those roles.

The concept of commitment has its basic referent in the networks of social relationships in which persons participate; as such, commitment is a social structural term. Associated with the ‘‘complex mosaic of differentiated parts’’ image of society is the recognition that persons conduct their lives not in the context of society as a whole but, rather, in the many contexts of relatively small and specialized social networks, networks made up of persons to whom they relate by virtue of occupancy of particular social positions and the playing of the associated roles. To say that persons are committed to some social network is to say that their relationships to the other members of that network depend on their playing particular roles and having particular identities: To the degree that one’s relationships to specific others depend on being a particular kind of person, one is committed to being that kind of person. Thus, commitment is measured by the costs of giving up meaningful relations to others should an alternative course of action be pursued. Commitment, so

defined and measured, is hypothesized by identity theory to be the source of the salience attached to given identities (Stryker 1968, 1980, 1987a).

Two analytically distinct and possibly independent dimensions or forms of commitment have been discerned: interactional and affective (Serpe 1987; Stryker 1968). The former has its referent in the number of relationships entered by virtue of having a given identity and by the ties across various networks of relationships (for instance, one may relate as husband not only to one’s spouse, her friends, and her relatives but also to members of a couples’ bridge club and other such groups). The latter has its referent in the depth of emotional attachment to particular sets of others.

Reciprocity among the three terms of the identity theory formula is again recognized; but the dominant thrust of the process is hypothesized to be as stated by the proposition, on the grounds that identity, as a strictly cognitive phenomenon, can change more readily than can commitment, whose conceptual core is interaction rather than cognition.

The empirical evidence brought to bear on the hypotheses contained in the fundamental identity theory formula has been supportive. Stryker and Serpe (1982) demonstrate that both time spent acting out a religious role and preferred distribution of time to that role are tied to the salience of the identity associated with the role; they demonstrate as well that the salience of the religious identity is tied to commitment (in this case, the measure of commitment combines interactional and affective commitment) to others known through religious activities. Burke and various associates (Burke and Hoelter 1988; Burke and Reitzes 1981; Burke and Tully 1977) show the link between identity and gender, academic attainment and aspirations, and occupational aspirations, finding evidence that the linkage reflects the commonality of meaning of identity and behavior. Lee (1998) finds that the correspondence of meanings of students’ personal identities and meanings they attach to those occupying positions in scientific disciplines predicts interest in science as well as appreciably accounting for gender differences in intention to become scientists. Serpe and Stryker (1987), using data on student-related identities obtained at three points in time from students

entering a residential college, provide evidence that the salience of these identities is reasonably stable over time; that in a situation in which earlier commitments have been attenuated by a move to a residential university, high identity salience leads to efforts to reconstruct social relationships that permit playing the role associated with the salient identity, efforts taking the form of joining appropriate organizations; and that when such efforts are not successful, the level of salience of the identity subsequently drops and self-structure is altered. Callero (1985), Callero and associates (1987), and Charng and associates (1988) show that commitment and identity salience add appreciably to the ability to account for the behavior of repeated blood donors. Sparks and Shepherd (1992) find, to their considerable surprise, that identity theory–based predictions stand up well in accounting for behavioral intentions with regard to green consumerism, the predicted relationships holding when examined in the context of the variables of a theory of planned behavior. Serpe (1987) shows that over time there is indeed a reciprocal relationship between commitment to various student role relationships and the salience of identities associated with those roles, and that the identity theory hypothesis arguing the greater impact over time of commitment on salience than vice versa is correct.

The success of identity theory attested to in this brief and incomplete review of empirical evidence notwithstanding, however, there is reason to believe that the theory requires development and extension beyond the basic proposition that has been the major focus of attention to this point. Indeed, such work has begun; and it has a variety of thrusts. How to incorporate varying degrees of situational constraint into the theory—the impact that variations in ‘‘choice’’ have on the ways in which the relationships among commitment, identity salience, and role performances play themselves out—is one such thrust (Serpe 1987). Another seeks to explore mechanisms underlying the linkages among commitment and identity salience, and identity salience and behavior; to this end, work (especially by Burke and Reitzes 1981, 1991) exploits the basic symbolic interactionist idea that it is commonality of meaning which makes social life possible. Some attention has been given to extending the applicability and predictive power

of identity theory by incorporating into it other than strictly role-based identities; in particular, the concern has been with what have been termed ‘‘master statuses’’ (such as age, gender, and class) and personal traits (such as aggressiveness and honesty), and the suggestion is that master statuses and traits may affect identity processes by modifying the meaning of the roles from which identities derive (Stryker 1987a). Stryker and Serpe (1994) treat the conceptual and measurement confusion of the importance ranking of identities and identity salience, in the process showing that both importance and salience respond to commitment, that importance and salience are related, and that both contribute to the prediction of role-related behavior.

An effort is being made to correct the almost totally cognitive focus of identity theory (as well as its parent and grandparent interactionist frameworks) by recognizing the importance of affect and emotion to the processes with which the theory is concerned. The earliest statement of the theory (Stryker 1968) posited a cathectic modality of self that parallels the cognitive modality from which the emphasis on identity flows; however, subsequent work on the theory has not pursued that idea. Stryker (1987b) has attempted to integrate emotion into the theory by arguing, with Hochschild (1979), that emotional expressions carry important messages from self and, beyond Hochschild, that the experiences of emotions are messages to self informing those who experience those emotions of the strength of commitments and the salience of identities.

Finally—and here work has barely begun—it is time to make good on the promise to provide more adequate conceptualization of the linkages between identity theory processes and the wider social structures within which these processes are embedded. From the point of view of structural symbolic interactionism, structures of class, ethnicity, age, gender, and so on operate as social boundaries making it more or less probable that particular persons will form interactional networks; in this way, such social structures enter identity theory directly through their impact on commitments. However, the relation of such structures to identity processes clearly goes beyond this direct impact; they affect not only the probabilities of

interaction but also the content (meanings) of the roles entailed in interaction and, thus, the meanings of identities, the symbolic and material resources available to those who enter interaction with others, and the objectives or ends to which interactions are oriented. Explication of these impacts, both direct and indirect, of social structure on the processes that relate commitment, identity salience, and role performance remains to be accomplished.

REFERENCES

Blumer, Herbert 1969 Symbolic Interactionism: Perspective and Method. Englewood Cliffs, N.J.: Prentice-Hall.

Burke, Peter J., and Viktor Gecas 1995 ‘‘Self and Identity.’’ In Karen Cook, Gary A. Fine, and James House, eds., Sociological Perspectives on Social Psychology. Needham Heights, Mass.: Allyn and Bacon.

Burke, Peter J., and Jon W. Hoelter 1988 ‘‘Identity and Sex–Race Differences in Educational and Occupational Aspirations Formation.’’ Social Science Research 17:29–47.

Burke, Peter J., and Donald C. Reitzes 1981 ‘‘The Link Between Identity and Role Performance.’’ Social Psychology Quarterly 44:83–92.

——— 1991 ‘‘An Identity Theory Approach to Commitment.’’ Social Psychology Quarterly 54:239–251.

Burke, Peter J., and Judy Tully 1977 ‘‘The Measurement of Role/Identity.’’ Social Forces 55:880–897.

Callero, Peter L. 1985 ‘‘Role-Identity Salience.’’ Social Psychology Quarterly 48:203–215.

———, Judith A. Howard, and Jane A. Piliavin 1987 ‘‘Helping Behavior as Role Behavior: Disclosing Social Structure and History in the Analysis of ProSocial Action.’’ Social Psychology Quarterly 50:247–256.

Charng, Hong-Wen, June Allyn Piliavin, and Peter L. Callero 1988 ‘‘Role-Identity and Reasoned Action in the Prediction of Repeated Behavior.’’ Social Psychology Quarterly 51:303–317.

Cooley, Charles H. (1902) 1983 Human Nature and Social Order. New Brunswick, N.J.: Transaction Books.

Hewitt, John P. 1997a Dilemmas of the American Self. Philadelphia: Temple University Press.

——— 1997b Self and Society, 7th ed. Needham Heights, Mass.: Allyn and Bacon.

Hochschild, Arlie R. 1979 ‘‘Emotion Work, Feeling Rules, and Social Structure.’’ American Journal of Sociology 85:551–575.

James, William (1890) 1990 Principles of Psychology. Chicago: Encyclopedia Britannica, Inc.

Lee, James D. 1998 ‘‘What Kids Can ‘Become’ Scientists? The Effects of Gender, Self-Concepts, and Perceptions of Scientists.’’ Social Psychology Quarterly

61:199–219.

MacKinnon, Neil J. 1994 Symbolic Interaction as Affect Control. Albany: State University of New York Press.

———, and Ann Statham 1985 ‘‘Symbolic Interactionism and Role Theory.’’ In Gardner Lindzey and Eliot Aronson, eds., The Handbook of Social Psychology, 3rd ed. New York: Random House.

Weigert, Andrew J. 1983 ‘‘Identity: Its Emergence Within Sociological Psychology.’’ Symbolic Interaction 6:183–206.

Markus, Hazel 1977 ‘‘Self-Schemas and Processing Information About the Self.’’ Journal of Personality and Social Psychology 35:63–78.

McCall, George, and J. S. Simmons 1978 Identities and Interaction, rev. ed. New York: Free Press.

Mead, George H. (1934) 1963 Mind, Self and Society. Chicago: University of Chicago Press.

Serpe, Richard T. 1987 ‘‘Stability and Change in Self: A Structural Symbolic Interactionist Explanation.’’ Social Psychology Quarterly 50:44–55.

———, and Sheldon Stryker 1987 ‘‘The Construction of Self and the Reconstruction of Social Relationships.’’ In Edward J. Lawler and Barry Markovsky, eds.,

Advances in Group Processes, vol. 4. Greenwich, Conn.: JAI Press.

Sparks, Paul, and Richard Shepherd 1992 ‘‘Self-Identity and the Theory of Planned Behavior: Assessing the Role of Identification with ‘Green Consumerism.’’’

Social Psychology Quarterly 55:388–399.

Stone, Gregory P. 1962 ‘‘Appearance and the Self.’’ In Arnold M. Rose, ed., Human Behavior and the Social Process. Boston: Houghton Mifflin.

Stryker, Sheldon 1968 ‘‘Identity Salience and Role Performance.’’ Journal of Marriage and the Family 30:558–564.

———1980 Symbolic Interactionism: A Social Structural Version. Menlo Park, Calif.: Benjamin/Cummings.

———1987a ‘‘Identity Theory: Developments and Extensions.’’ In Krysia Yardley and Terry Honess, eds.,

Self and Society: Psychosocial Perspectives. New York: Wiley.

———1987b ‘‘The Interplay of Affect and Identity: Exploring the Relationships of Social Structure, Social Interaction, Self and Emotion.’’ Paper presented at annual meeting of the American Sociological Association, Chicago.

———, and Richard T. Serpe 1982 ‘‘Commitment, Identity Salience, and Role Behavior: Theory and Research Example.’’ In William Ickes and Eric S. Knowles, eds., Personality, Roles, and Social Behavior. New York: Springer-Verlag.

———, and Richard T. Serpe 1994 ‘‘Identity Salience and Psychological Centrality: Equivalent, Overlapping, or Complementary Concepts?’’ Social Psychology Quarterly 57:16–35.

SHELDON STRYKER

ILLEGAL ALIENS/

UNDOCUMENTED

IMMIGRANTS

See International Migration.

ILLEGITIMACY

Until the 1960s, it was widely assumed that marriage was a universal or nearly universal institution for licensing parenthood. Marriage assigned paternity rights to fathers (and their families) and guaranteed social recognition and economic support to mothers and their offspring. According to Malinowski (1930), who first articulated ‘‘the principle of legitimacy,’’ and to Davis (1939, 1949), who extended Malinowski’s theory into sociology, marriage provides the added benefit to children of connecting them to a wider network of adults who have a stake in their long-term development.

This functional explanation for the universality of marriage as a mechanism for legitimating parenthood became a source of intense debate in anthropology and sociology in the 1960s. Evidence accumulated from cross-cultural investigations showed considerable variation in marriage forms and differing levels of commitment to the norm of legitimacy (Bell and Vogel 1968; Blake 1961; Coser 1964; Goode 1961). More recently, historical evidence indicates that the institution of marriage was not firmly in place in parts of Western Europe until the end of the Middle Ages (Glendon 1989; Goody 1983).

The accumulation of contradictory data led Goode (1960, 1971) to modify Malinowski’s theory to take account of high rates of informal unions

and nonmarital childbearing in many New World nations and among dispossessed cultural minorities. Goode (1971) argued that the norm of legitimacy was likely to be enforced only when fathers possessed wealth and property or when their potential economic investment in child rearing was high. Therefore, he predicted that when Agiving a name’’ to children offers few material, social, or cultural benefits, the norms upholding marriage will become attenuated.

So vast have been the changes in the perceived benefits of marriage since the 1960s in the United States and most Western nations that even Goode’s modification of Malinowski’s theory of legitimacy now seems to be in doubt (Cherlin 1992; Davis 1985). Indeed, the term ‘‘illegitimacy’’ has fallen into disfavor precisely because it implies inferior status to children born out of wedlock. Both legal and feminist scholars have been critical of the notion that the presence of a father confers status on the child (Burns and Scott 1994; Mason et al. 1998). The nuclear unit (biological parents and their offspring)—once regarded as the cornerstone of our kinship system—remains the model family form, but it no longer represents the exclusive cultural ideal, as was the case in the mid-1960s. The incentives for marriage in the event of premarital pregnancy have declined, and the sanctions against remaining single have diminished (Cherlin 1999; McLanahan and Casper 1995). In the 1990s, considerable scholarly attention and public policy debate has been devoted to ways of restoring and reinvesting in the institution of marriage (Furstenberg 1996; McLanahan and Sandefur 1994; Popenoe et al. 1996).

TRENDS IN NONMARITAL CHILDBEARING

Premarital pregnancy has never been rare in the United States or in most Western European nations (Burns and Scott 1994; Goode 1961; Smith 1978; Vinovskis 1988). Apparently, the tolerance for pregnancy before marriage has varied over time and varies geographically at any given time. Throughout the first half of the twentieth century, premarital pregnancy almost always led to hasty marriages rather than out-of-wedlock births—even for very young women (O’Connell and Moore 1981; Vincent 1961). In 1940, illegitimacy was

uncommon in the United States, at least among whites. Nonmarital births were estimated at about 3.6 per 1,000 unmarried white women, while the comparable rate for nonwhites was 35.6. For all age groups, among whites and nonwhites alike, a spectacular rise occurred over the next five decades (Clague and Ventura 1968; Cutright 1972; McLanahan and Casper 1995).

In the 1960s and the 1970s, nonmarital childbearing rates continued to increase for younger women, albeit at a slower pace, while for women in their late twenties and thirties rates temporarily declined. Then, in the late 1970, nonmarital childbearing rates rose again for all age groups and among both whites and African Americans. This rise continued until the mid-1990s, when levels of nonmarital childbearing stabilized or even declined (Ventura et al. 1996). Since the early 1970s, rates of marriage and marital childbearing have fallen precipitously. Thus, the ratio of total births to single women has climbed continuously (Smith et al. 1996). Nearly a third of all births (32.4 percent) in 1996 occurred out of wedlock, more than seven times the proportion in 1955 (4.5 percent) and more than twice that in 1975 (14.3 percent). The declining connection between marriage and parenthood is evident among all age groups but is especially pronounced among women in their teens and early twenties. Three out of four births to teens and nearly half of all births to women ages twenty to twenty-four occurred out of wedlock. Virtually all younger blacks who had children in 1995 (more than 95 percent) were unmarried, while two-thirds of white teens and more than a third of white women twenty to twenty-four were single when they gave birth.

Nonmarital childbearing was initially defined as a problem among teenagers and black women (Furstenberg 1991). But these recent trends strongly suggest that disintegration of the norm of legitimacy has spread to all segments of the population. First the link between marriage and sexual initiation dissolved, and now the link between marriage and parenthood has become weak. Whether this trend is temporary or a more permanent feature of the Western family system is not known. But public opinion data suggest that a high proportion of the population finds single parenthood acceptable. A Roper study (‘‘Virginia Slims American

Women"s Opinion Poll’’ 1985) revealed that 49 percent of women agreed that ‘‘There is no reason why single women should not have children and raise them if they want to.’’

Citing similar attitudinal evidence from the National Survey of Families and Households in 1987–1988, Bumpass (1990) concludes that there has been an ‘‘erosion of norms’’ proscribing nonmarital childbearing. He concludes that this behavior is not so much motivated by the desire to have children out of wedlock as it is by the reduced commitment to marriage and the limited sanctions forbidding nonmarital childbearing. Bumpass argues that much of the nonmarital childbearing is unplanned and ill timed.

THE CONSEQUENCES OF NONMARITAL

CHILDBEARING

Although extensive research exists on the economic, social, and psychological sequelae of single parenthood for adults and children, relatively little of this research has distinguished between the consequences of marital disruption and nonmarriage (Furstenberg 1989; Furstenberg and Cherlin 1991; Garfinkel and McLanahan 1986; Maynard 1997). A substantial literature exists on the consequences of nonmarital childbearing, but it is almost entirely restricted to teenage childbearers (Chilman 1983; Hofferth and Hayes 1987; Institute of Medicine 1995; Miller and Moore 1990; Moore et al. 1986). It is difficult, then, to sort out the separate effects of premature parenthood, marital disruption, and out-of-wedlock childbearing on parents and their offspring.

Nonmarital childbearing most certainly places mothers and their children at risk of long-term economic disadvantage (Institute of Medicine 1995; Maynard 1997; McLanahan and Booth 1989). Out- of-wedlock childbearing increases the odds of going on welfare and of long-term welfare dependency (Duncan and Hoffman 1990). The link between nonmarital childbearing and poverty can probably be traced to two separate sources. The first is ‘‘selective recruitment,’’ that is, women who bear children out of wedlock have poor economic prospects before they become pregnant, and their willingness to bear a child out of wedlock may reflect the bleak future prospects of many unmarried pregnant women, especially younger women

(Furstenberg 1990; Geronimus 1987; Hayes 1987; Hogan and Kitagawa 1985; Maynard 1997). But is also likely that out-of-wedlock childbearing—par- ticularly when it occurs early in life—directly contributes to economic vulnerability because it reduces educational attainment and may limit a young woman’s prospects of entering a stable union (Furstenberg 1991; Hofferth and Hayes 1987; Hoffman et al. 1993; Trussell 1988).

If nonmarital childbearing increases the risk of lengthy periods of poverty for women and their children, it is also likely that it restricts the opportunities for intraand intergenerational mobility of families formed as single-parent units. Growing up in poverty restricts access to health, high-quali- ty schools, and community resources that may promote success in later life (Ellwood 1988; Wilson 1987). Apart from the risks associated with poverty, some studies have shown that growing up in a single-parent family may put children at greater risk because they receive less parental supervision and support (Amato and Booth 1997; Dornbush 1989; McLanahan and Booth 1989; McLanahan and Sandefur 1994). As yet, however, researchers have not carefully distinguished between the separate sources of disadvantage that may be tied to nonmarital childbearing: economic disadvantage (which could restrict social opportunities or increase social isolation) and psychological disadvantage (which could foster poor parenting practices or limit family support).

Even though nonmarital childbearing may put children at risk of long-term disadvantage, it is also possible that over time the advantages conferred by marriage may be decreasing in those segments of the population that experience high rates of marital disruption (Bumpass 1990; Edin 1998; Furstenberg 1995). Moreover, the social and legal stigmata once associated with nonmarital childbearing have all but disappeared in the United States and many other Western nations (Glendon 1989). Over time, then, the hazards associated with nonmarital childbearing (compared with illtimed marital childbearing) for women and their children may have declined. Whatever the reasons for these trends, it appears that nonmarital childbearing may have peaked by the mid-1990s. Whether the leveling off in rates of nonmarital childbearing signals a shift in family formation patterns or is

merely a response to the robust economy of the 1990s remains to be seen.

NONMARITAL CHILDBEARING AND

PUBLIC POLICY

Growing rates of nonmarital childbearing in the United States and many other Western nations suggest the possibility that the pattern of childbearing before marriage or between marriages may be spreading upward into the middle class. In Scandinavia, where marriage has declined most dramatically, it is difficult to discern whether formal matrimony is being replaced by a de facto system of informal marriage (Hoem and Hoem 1988). If this were to happen, the impact on the kinship system or the circumstances of children might not be as dramatic as some have speculated. But if the institution of marriage is in serious decline, then we may be in the midst of a major transformation in the Western family.

The weakening of marriage has created confusion and dispute over parenting rights and responsibilities. A growing body of evidence indicates that most nonresidential biological fathers, especially those who never marry, typically become disengaged from their children (Arendell 1995; King and Heard 1999; Seltzer 1991; Teachman 1990). Most are unwilling or unable to pay regular child support, and relatively few have constant relationships with their children. Instead, the costs of child rearing have been largely assumed by mothers and their families, aided by public assistance. A minority of fathers do manage to fulfill economic and social obligations, and some argue that many others would do so if they had the means and social support for continuing a relationship with their children (Marsiglio 1998; Smollar and Ooms 1987).

The uncertain relationship between biological fathers and their children has created a demand for public policies to shore up the family system (Garfinkel et al. 1996; Popenoe 1996). Widespread disagreement exists over specific policies for addressing current problems. Advocates who accept the current reality of high levels of nonmarriage and marital instability propose more generous economic allowances and extensive social support to women and their children to offset the limited

economic role of men in disadvantaged families (Ellwood 1988). Critics of this approach contend that such policies may further erode the marriage system (Vinovskis and Chase-Lansdale 1987). Yet few realistic measures have been advanced for strengthening the institution of marriage (Furstenberg and Cherline 1991).

Enforcement of child support has attracted broad public support. A series of legislative initiatives culminating in the Family Support Act of 1988 have increased the role of federal and state governments in collecting child support from absent parents (typically fathers) and standardizing levels of child support. There has been a steady but modest improvement in the collection of child support in the 1990s. It is much less clear whether the strengthening of child support has worked to the benefit of children (Furstenberg et al. 1992; Garfinkel et al. 1996). It is too early to tell whether these sweeping measures will succeed in strengthening the economic contributions of fathers who live apart from their children. And, if it does, will greater economic support by absent parents reinforce social and psychological bonds to their children (Furstenberg 1989; Garfinkel and McLanahan 1990)? The other great experiment of the 1990s was the Welfare Reform Act in 1996, which replaced the longstanding entitlements to public assistance with temporary provisional support. It is much too early to tell what, if any, the effects of this policy will be on marriage and fertility practices. Advocates of welfare reform claimed that it would reduce out-of-wedlock childbearing and help restore marriage (Murray 1984). But the link between welfare payments and marriage patterns has never been strong (Moffitt 1998). Still, it may be possible to devise a test of the consequences of the different policies given the large state variations in program implementation.

As for the future of marriage, few, if any, sociologists and demographers are predicting a return to the status quo or a restoration of the norm of legitimacy. Short of a strong ideological swing favoring marriage and condemning nonmarital sexual activity and childbearing, it is difficult to foresee a sharp reversal in present trends (Blankenhorn et al. 1990). Predicting the future, however, has never been a strong point of demographic and sociological research.