Social Dynamics of Adolescent Fertility in Sub-Saharan Africa (1993)

As sub-Saharan Africa approaches the year 2000, its fertility rates remain the highest in the world. Increases in contraceptive use and decreases in fertility have been observed recently in Kenya, Botswana, and Zimbabwe (Cohen, 1993; Working Group on Factors Affecting Contraceptive Use, 1993). Most family planning programs, however, encounter reactions ranging from polite interest to resentment.¹ Why are some of these reactions so strong?

Virtually everyone in Africa, from rural farmer to sophisticated urbanite, experiences intense pressure to become a parent. As we have seen, marriage is quite important in sub-Saharan Africa. Yet the process of maturation is geared less toward marriage than toward preparation for parent-hood. Fortes (1978:121) was quite right when he pointed out that "... it is not marriage but parenthood that is the primary value associated with the family in West Africa." Throughout the subcontinent, failure to bear a child has long been a major cause of divorce or of men marrying additional wives (for example, Chojnacka, 1980; Denga, 1982; Brandon, 1991).

Although fertility is intensely valued, two powerful conditions are attached to it. First, it ideally occurs within "sanctioned" states. Although

outsiders often assume that this sanctioned state is marriage, entitlement to reproduce is more fundamentally earned by undergoing ritual or learning experiences that prepare individuals to be responsible parents. The second condition for sanctioned reproduction is that the child have an acknowledging social father, whether or not he is the actual biological father: someone who will share with the mother the financial responsibility of raising the child, whose own family will enlarge the network of potential allies and supporters for the woman's family, and who may offer spiritual guidance that is essential to the child's development.

As Chapter 3 explains, a premarital birth, the variable that this report has taken as its initial focus, is not necessarily unsanctioned, so long as the father acknowledges his paternity. We therefore treat an increase in pre-marital births as a possible indicator of a more serious change: a rise in unsanctioned births that occur either without paternal recognition, or outside of what is considered the proper life cycle phase. Because quantitatively measuring such changes is virtually impossible given the data now available, this chapter lays out a broader historical and cultural perspective that may help to illuminate these aspects of reproduction.

Focusing on reports of precolonial dynamics, we stress that sexual expression and legitimate reproduction were two conceptually separate activities. In many African societies young people were allowed, even encouraged, to engage in limited kinds of sexual expression; but they were not supposed to reproduce until qualified, whether through training (that is, formal education or informal apprenticeship) or ritual initiation, for the labor-intensive and costly process of raising children. There were breaches of this norm, no doubt; many references to the past both by ethnographers and by societal elders may be idealized statements. Still, the overwhelming perception today is that family elders in Africa are losing control over youthful reproductive life and practices. The perceived results are both the rupture of a carefully ordered sequence of maturational events that under-laid successful reproduction, and, more recently, a decline in men's willingness to acknowledge paternity. The ultimate result is an increase in unsanctioned births.

The importance of the events that cluster around adolescence in Africa can be comprehended within a broader socioeconomic framework that is characteristic especially of the past. In this framework wealth and value derived less from land, as in Europe, than from human capital. (See Goody, 1976, for the major statement of this contrast.) At low population densities the conquest of land could bring neither wealth nor power. Political com-

munities and productive economies could be built up only by a deeply attentive construction of ties of personal dependence among people. The ethnographic and historical literature has characterized these social dynamics as a "wealth in people" system, in which the control of dependents meant status and security (for variants on this theme, see Mair, 1953; Fallers, 1964; Goody, 1971; Meillassoux, 1975; Kopytoff and Miers, 1977; Bledsoe, 1980; Etienne, 1983).

Although strategies involving the derivation of wealth from people could be applied to all relationships, they were deployed most fully with respect to marriage, reproduction, and child raising. The ethnographic literature suggests strongly that these societies encouraged prolific childbearing and that people sought in various ways to gain control over women's reproductive capacities. Men sought to maximize their wealth by taking more than one wife, and by choosing young women, who had a longer reproductive life remaining to them than did women in their twenties and thirties. Discussing "maidenhood," Whiting et al. point out that African societies attempt to "make full use of the reproductive lives of their females" (1986:290). Polygynous marriage and the welding of the complex kinship ties it created were among the most important strategies a man could employ to build a prestigious personal career. Marriage started a man's autonomous career of building up his own "wealth in people," and it followed a clear accumulative dynamic. Wives brought with them skills in farming and advantageous political connections, and they bore children who themselves could eventually contribute to the family's name and strength.

Marriage and parenthood not only contributed to a man's status; they were measures of it. For individual men, multiple wives and sexual partners were both an avenue to wealth and a public expression of power. They were also a means of creating the next generation and endowing it, in turn, with the resources needed to survive childhood and to be sponsored to full maturity. The lack of a wife, and by extension the lack of full sexual opportunity, was a sign of youth, all levels of servility, and poverty. Women's sexual demands, by contrast, were not related to their power in the larger political arena or to the accumulation of wealth. Polyandry occurred in some places, but it was not available for all women (Douglas, 1963). Where women had publicly acknowledged multiple partners, the practice was uncommon or sequential, at least where sex was not an income-earning occupation, as it became during the colonial period (see Cohen, 1969; White, 1990).

Especially in the past, having children was the essential mark of manhood and womanhood (for Nigeria, see Isichei, 1978). The Tallensi of

northern Ghana believed that "no one is so unfortunate as a woman who has no children" (Fortes, 1959:37). In Kenya the Meru people forbade childless men to participate in certain ceremonies and rituals (Mwambia, 1973). In eastern and southern Africa a barren woman had failed to fulfill her obligation in life; her condition was viewed as a punishment from the gods or the ancestors or as the result of witchcraft (Kershaw, 1973; Mwambia, 1973; Ngubane, 1977). Many societies buried childless women with minimal respect if not with contempt: Their bodies were sometimes mutilated (for the Yoruba of Nigeria, see Caldwell and Caldwell, 1985; for the Gusii of Kenya, see Mayer, 1973).

That societies of the past emphasized fertility should not be surprising. Yet the desire for high fertility persists as a strong value even in contemporary urban situations, in which the utility of children as subsistence laborers has diminished and their costs of maintenance and education have increased.

Children still play important roles in the domestic economy. Girls as young as 5 or 6 help with cooking, cleaning the house, bringing water, looking after smaller children, and, at times, trading. Though they have more license to roam about with friends, boys, too, are put to work. As adults, children who go elsewhere to obtain employment are reminded at every opportunity that their remittances are vital to their kin. Those who remain take over the household and its financial matters, allowing parents to assume the roles of elder statesmen who bestow advice and blessings on the young.

For men, children forge firmer links to potentially advantageous in-laws. They also enhance men's long-term hopes for improving the material well-being of their families and their ability to compete with rival families for positions of political leadership. In Freetown, Sierra Leone, men are warned against marrying infertile women, and some prefer proof of fertility before marriage (Harrell-Bond, 1975; see also Obbo, 1987, for Uganda, and Mbatha, 1983, for Swaziland). The Demographic and Health Survey for Liberia asked women if men preferred to marry a woman who had already given birth; 27 percent responded positively, 36 responded negatively, and 36 percent said that they did not know.

Women's needs for children are no less acute. A married woman needs children to justify demands on her husband's wealth and estate. Co-wives feel this need most painfully. No matter how well they get along, polygynous wives jealously observe how many children the other wives bear; a subfertile wife must watch her productive labor for the household going to benefit her co-wives' children. Fearing to short-change any of his wives of children, a polygynous man is a reluctant advocate of birth control (Bledsoe, 1992). Even for unmarried women, children are valued. In Sierra Leone, the changes wrought by Christianity and education have not dampened the attitude that unmarried mothers are far better off than childless women.

Although they prefer their daughters to marry, desperate parents may pressure an older educated daughter to have children, though no permanent husband is in sight.

In western Nigeria, parents now expect little daily material support from adolescent and adult children or their spouses (Guyer, 1988a). Farm work has become increasingly commercialized, and fathers cannot count on their offspring to work from a sense of obligation (see also Berry, 1985). Sons can be too busy elsewhere in the local economy, or they can disappear into the urban labor market. But although adults have fewer sanctions to impose on their sons and daughters themselves, the socioeconomic context still offers indirect benefits to bearing children. Economic instability pressures women to establish immediate lateral links with different men and their resource networks. Lateral strategies can yield dividends in less time than it takes children to mature, and they offer greater breadth and flexibility of networks. At the heart of these lateral strategies are children. A woman can press her demands on a man, whether or not they call their relationship a marriage, with far greater leverage if she has a child by him—a strategy Guyer calls ''polyandrous motherhood.'' Marriage, in fact, can become almost incidental to a woman's reproductive career. "The child is the key; without it there is no basis to claim anything beyond the moment of the relationship" (Guyer, 1988b; see also Karanja, 1987).

In many African countries, successful children become brokers for their families in broader national contexts of patron-clientelism. Jobs, scholarships, money, and land are dispensed through personal ties to powerful mediators who can maneuver within government bureaucracies to obtain resources for their dependents. With precipitous declines in national economies, people have an even greater need for patrons well connected to the urban and government bureaucracies to help them bypass cumbersome channels during shortages, and to provide them with crucial ties to the international world for travel, jobs, and access to hard foreign currency.

For precolonial Africa, temporally ordered events and the predictability they lent to life were fundamental to society. Social rankings were often based on temporal ordering: In parts of eastern and southern Africa, siblings were ranked by birth order, co-wives by sequence of marriage, generations through age grades, and kinship groups through the reputed order of their ancestors' order of birth or arrival into the area. A temporal ordering and pacing of events was also important to the life cycle of individuals. Rather than a passively experienced series of maturational events, life was cast as a process of ordered achievements, a succession of ritualized steps (see, for example, Menkiti, 1979; Ottenberg, 1989). Even as adults, people

might still be working up the gradations of title societies whose final stages conferred full social majority.

It was particularly important to confine conception and childbearing to a properly ordered and paced life sequence from birth, through puberty, training, marriage, parenthood, entry into work, and death. Obviously, the preferred sequence of life events varied from one society to another. And the sequence was not intractable; it could be followed rigorously or adapted to new circumstances. Yet when a breach in the sequence was perceived, condemnation could be swift. One of the most important pieces of evidence that fertility was not always valued comes from widespread reports of indigenous contraceptives and abortifacients (and even, occasionally, infanticide) among unmarried women and among married women whose husbands had been away (see Caldwell and Caldwell, 1988; Lema and Kabeberi-Macharia, 1990, for Kenya; Mueller, 1976, for the Rukuba in Nigeria; and Schapera, 1933, for the Tswana of South Africa).

Because many of these crucial events occurred around adolescence, the way adolescence proceeded was considered vital to a successful reproductive life, and thus to the well-being of the family and community. Adolescence set the stage for producing the next generation, whose skills and character would ensure continuity and security in an environment fraught with uncertainty and risk:

Scholars have tended to emphasize the benefits that adults reaped through children. But adults also had to invest considerable material and social capital in the young so they could meet their full potentials and assist their families to progress. The need to ensure societal continuity was deemed so important that the task of training and sponsoring a child could not be trusted to the parents alone; it was instead considered a societal responsibility (Sala-Diakanda, 1991).

The cultural emphasis placed on raising the young made socialization a costly, labor-intensive process in precolonial Africa. The sequence of events making up the reproductive cycle of procreation, socialization, and social attachment was minutely specified, lavishly supported by material and hu-

man resources, and attentively sanctioned by religious and political powers. To begin a social career, a child established his position in his parents' kin groups typically by completing specified ritual or economic transactions. A key element of this process was the consolidation of his relationships with an expanding array of individuals and groups; these relationships were expected to be reaffirmed repeatedly at key points in his development.

The Igbo of eastern Nigeria offer a striking example. Writing of the 1950s, Ottenberg (1989) reports that by the time a young man was about 18 he had been through two major rituals of infancy and early childhood, had belonged to three successive boyhood societies, each with its initiation rituals, and had finally graduated in a formal ceremony to membership in the adult male society. He had been trained as a warrior and had acquired basic productive skills and religious knowledge. He had been allowed an indulgent, but not usually sexual, friendship with a girl from his peer group in early adolescence, and had been betrothed for a marriage that would not take place until he was considered fully mature, with his own house and gardens, in his late twenties. Every stage from conception to marriage involved making or acknowledging ever-widening ties with peers, sponsors, teachers, patrons, and future in-laws. Each stage had to be correctly enacted for the young man's own social and reproductive career to be successful. Deviations and failures could bring large fines, costly redemptive ceremonies, or physical punishment.

After initiation, a young man faced a period of up to 10 years when sexual access was "uncertain and erratic" (Ottenberg, 1989:301) because the only possible partners were the few unmarried girls, divorced women, stranger-prostitutes, and married women bold enough to engage in adultery. The passage of boyhood events and their concomitant buildup of carefully forged relationships gradually began to merge with overtures toward a young woman's family by the young man's family that signified the beginning of a conjugal relationship. Marriage itself was therefore a gradual achievement of status. Like the process of moving from one maturational phase to another, each step toward an eventual conjugal tie might be symbolized separately by the long series of exchanges. Ngoi (1950) gives an early participant's account for the Nkundu of Zaire: Three differently named sets of gifts transferred different rights and obligations between the two families, and each of the nine components of the last gift recognized the particular content of the relationship of the gift's recipient to the bride.

Throughout Africa, societies followed various versions of this kind of elaborated sequence of maturation. In eastern and southern Africa, where age-grade organizations were important elements of political organization, young men graduated from the boyhood groups into military regiments before taking up marriage and reproduction. In western and Central Africa the rituals of initiation took on heightened drama: Young people of both

sexes could spend months or even years traversing all the stages of the religious threshold into adulthood (see, for example, MacCormack, 1982). Especially for young women, these stages often blended with those ushering in marriage. So intense was this emphasis on proper preparation for reproduction that in many cases it strongly affected the interpretations local people placed on modern innovations. Among the Bangangte of Cameroon, Feldman-Savelsberg (1989) reports, in the past a girl was secluded around puberty and fed rich foods to fill out her frame, a widespread western African practice that MacCormack (1982) argues may have accelerated menarche. Many older Bangangte women equate modern schooling literally with this practice: "Girls [are] sitting in school doing no physical labor, eating up household resources, and preparing themselves for marriage" (Feldman-Savelsberg, 1989:216).

This emphasis on taking time to develop the potential in children through ritual and training meant that no matter whether young men and women were physiologically mature enough to reproduce, they were expected to be prepared socially for the vital role of reproducing and supporting the next generation. This expectation, in turn, implied that sexual access and, to a much greater extent, parenthood were to be earned; one became entitled to reproduce. During the various stages of socialization, children and adolescents were not supposed to be endowed with procreative capacities. Robertson (1984:167) quotes an older Fanti woman in Ghana who describes her mother's reaction to her menarche in 1928:

Only individuals who had achieved social adulthood were considered able to generate the recognized value and wealth that bearing and raising children demanded. As Mutambirwa (1990:4) explains for Zimbabwe, the milestone of maturity was reached when a young couple could "deny themselves of physical comforts in order to promote the physical, social and spiritual well being of their. . . child . . . ".

Where unmarried men were allowed certain kinds of sexual access they could not usually claim the children (see, for example, Laburthe-Tolra, 1981, for the precolonial Beti of Cameroon). To be allowed to impregnate but not to claim the child was considered "working for nothing." In the extreme case of this philosophy, low-status people of all kinds, clients, youths, and slaves could be forbidden ever to work directly on any goods that counted as wealth, let alone claim them, so clear was the association between status and the right to generate wealth, of which children were an important component (Mandala, 1990).

Underlying this approach to life was a highly positive attitude to sexual relations, especially for men. Yet sexual access came with complex directives. Although many African societies allowed and even encouraged young men to be sexually expressive, they were not necessarily entitled yet to procreate. Sexual expression might precede marriage, but it was held in check by various means until reproduction could legitimately ensue. Barriers sharply separated sexuality from legitimate reproduction even in practices involving women of low status. In a historical study of the now-defunct practice of pawnship—giving a child temporarily as security for a loan, Falola (1993) reports that in pre-twentieth-century Nigeria, even though a female pawn was subservient, the contract explicitly forbade sexual access to her on pain of a fine.

Attaining a state in which reproduction was legitimate almost always entailed ritual passage of some sort through learning or testing of mettle. To attain legitimate reproductive status, girls often underwent a series of expensive initiation ceremonies. In some instances initiation included female circumcision (more accurately, clitoridectomy), a practice that arouses considerable controversy. As practiced in the women's secret Sande society of Sierra Leone and Liberia, it has been explained as a measure to render a girl wholly female (see, for example, MacCormack, 1975), to facilitate sexual relations (Sawyer and Todd, 1970), or to cool the sexual passions of women and curb their infidelity (Schwab, 1947). Speculating about Nigeria, for example, Susan Rich (personal communication) posits an inverse relation between early marriage and female circumcision. Societies with very early marriage are less likely to circumcise girls because the problem of premarital births is circumvented.

Even in most matrilineal societies, in which the child's membership in a lineage was not at stake, it was crucial for the mother to have already passed through the puberty ceremonies that legitimated her capacity to pro-create (see, for example, Richards, 1982, on the Bemba of Zambia; Brydon, 1987, on the Avatime of Ghana; Fortes, 1950, on the Ashanti). Compared to patrilineal groups, matrilineal groups such as the Akan of Ghana appeared to be more tolerant of adolescent sexual activity and births (for the Coniagui of Guinea, see Gessain, 1971; for the Ashanti of Ghana, see Fortes, 1978). Even so, among the Ashanti in the 1930s and 1940s it was a serious crime for a girl to become pregnant before her nubility ceremony (Fortes, 1954:265). According to Fortes, who described "the religious horror of pregnancy before the nubility ceremony," "the shame for the girl's parents is so great that, as the Ashanti say, they would rather be dead than endure it" (cited in Caldwell and Caldwell, 1988). Similarly, Richards (1982:33) observed that among the matrilineal Bemba of Zambia, "[o]ne of the most

fatal acts of all was for a girl to bear a child before she had been initiated .... Her child would then be a creature of ill-omen ... who would bring misfortune on any village in which it lived" (see also, for example, Gessain, 1971, for Guinea).

In many areas these ritual boundaries remain so important that some societies, both patrilineal and matrilineal, have deliberately reduced the age of these rituals to prevent the otherwise unacceptable procreation that occurs under the new conditions of the twentieth century (see Ottenberg, 1989, on circumcision of Igbo boys in Nigeria; and Bledsoe, unpublished field notes, on initiations of Mende girls in Sierra Leone). With the same aim, other societies have mandated the retroactive power of the rituals (see Brydon, 1987, on the adulthood ceremonies of Avatime girls in Ghana). In a contemporary variant on this theme, in urban Accra marriage is no longer a requisite for women's entry into socially defined adulthood; but initiation—though it may now take place after the births of several children—is still considered mandatory (Brydon, 1983).

Even within marriage there were strictures about when reproduction could, and could not, legitimately occur. The ubiquitous requirement to observe a period of abstinence after the birth of a child is an example (Page and Lesthaeghe, 1981). Other constraints within marriage applied in western African societies in which women marry at very early ages. Such sanctions characterized especially those societies influenced by Islamic doctrine, which employed early marriage to assuage worries about female sexual purity and unsanctioned births (see Dupire, 1963, for the Fulani in Niger, and Schildkrout, 1983, for the Hausa in northern Nigeria). Although the young bride might live in the husband's household and perform domestic chores, full sexual relations were sometimes delayed for several months or even years so that she could mature physically, as Marris (1962) indicates for the Hausa. The senior wife was often the one to inform her husband when the young woman was old enough to begin sexual relations and childbearing, an event that was frequently marked ritually. In other cases, a girl as young as 12 might be married, according to her society's customs but remain with her own parents for some years until she matured. "Though ... the couple are considered man and wife, ... several months and sometimes up to several years will go by before the next ceremony ... (fetching the bride) ... occurs" (Riesman, 1992:84, for the Fulani, in what is now Burkina Faso).

The Senegalese fertility survey found quantitative evidence for this pattern among the Fulani people (Republique du Sénégal, 1981). It pointed out that Fulani women have a very low median age at marriage, 15.4 years, about two years younger than women in the country's other ethnic groups. Yet they tend to be less fertile during the first years of marriage than women in other groups: 138 births per thousand women in the first five

years of marriage, compared with 159 for the Wolof, 177 for the Mandinka, and so on. This delay could well be attributed to physical immaturity, and a practice referred to as jebalé (Republique du Sénégal, 1981). According to Pierre Ngom (personal communication), this is a Wolof loan word, a verb meaning "to provide." It refers to the parents' act of finally "providing" the man with the wife he married some time ago. In the meanwhile, the bride might live with her parents, while the husband paid bridewealth installments and she matured.

The following excerpts from a woman's life history in rural Sierra Leone vividly highlight some analogous patterns in a different cultural setting (Bledsoe, unpublished field notes, 1982). The case stemmed from a healing episode, in which an Islamic holy man was said to have cured a barren woman's sterility. In gratitude, the woman's husband sent the baby daughter she bore to be the holy man's wife. As told by the now-adult daughter, this life story reveals the considerable danger in taking Western categories of "marriage" and "wife" at face value in Africa. The woman begins from her early childhood:

The preceding discussion has established the importance of ritual or economic preparedness for childbearing. The second major stipulation that

most African societies made about reproduction was that a man should be willing to undertake the social and economic responsibilities of fatherhood. Although this requirement sounds very much like Western notions of legitimacy, there are two acute differences. First, to many Westerners, legitimacy rests largely on whether the mother's husband is the biological father of the child she bears. Second, and more important, Westerners have been preoccupied with timing: whether the conception, but most especially the birth, occurred while the mother was married, with divorce or widowhood being the main allowable exceptions.² Worries about inheritance of small, finite supplies of land or property may have prompted European societies to create a "highly developed concept of illegitimacy" (Laslett, 1980:9). In any event, European societies also seemed "to have classed the largest possible number of children as illegitimate" (Laslett, 1980:9).

Most African societies see the matter quite differently. Legitimacy is not always linked to being born within the temporal confines of a recognized union, nor is it tightly linked to biological descent. Indeed, although biological and social fathers should ideally be the same, the former are usually recognized as less important to the welfare of the child than the latter (see, for example, Fortes, 1950). Whether the parents have actually concluded a marriage is considerably less important than whether a man is willing to acknowledge fatherhood and to claim the social and economic responsibilities of that role.

The reason why securing a responsible father is so important to a child is not only that the father will share the financial responsibility of raising the child, but also that his family will enlarge the child's network of future allies and supporters. None of the vital life events that an individual undergoes—birth, education, marriage, childbearing, or death—should be accomplished in the absence of supportive kin or friends. Ties with other families and groups are needed to support and affirm the important events in life. The corollary of this observation is that vital events such as births, in turn, should ideally lead to the construction of newer, more solid ties with people outside the immediate family, a point that Riesman eloquently articulates in his study of Burkina Faso (1992). Childbearing, like other major events, should create new ties and deepen existing ones.

Especially in the past, having a recognized father located a child securely within the protection of two kin groups. In patrilineal groups, the

most prevalent in Africa, paternity gave a young man a name, an identity, and a source of wealth for marriage and inheritance. In matrilineal groups, children were the basis of the woman's lineage strength and paternal legitimacy per se was less important than it was among patrilineal groups. Yet even in these groups, an acknowledged paternity was vital, as Fortes (1950:266) confirmed:

Because of their overall value, children born to an adulterous union might be assimilated into either their mothers' or their fathers' families. But assimilation was not automatic. Such births might necessitate the official identification of the father and expensive redress of some kind, whether indemnities or rituals of purification (Ottenberg, 1989). Children lacking acknowledging fathers might be permanently assigned to servility (Laburthe-Tolra, 1981, on the Beti of Cameroon). The possibility that the famous Shaka, founder of the Zulu Empire in the nineteenth century, was conceived under unacceptable circumstances—and suffered the consequent severe denigration in childhood—was used by novelists, historians, and local people themselves to explain his later ruthlessness (see, for example, Mofolo, 1931).

Just as fears of reproduction during unsanctioned states made people devise creative protective measures for young women, so was young men's behavior channeled. In eastern and southern African societies, adolescent boys were taught how to have sexual relations between the thighs of their girlfriends, so as to prevent the pregnancy that was the sole prerogative of the girls' eventual husbands, older and more powerful men, to initiate. Among the Kikuyu of Kenya such practices were called ngweko, and both girls and boys were instructed in their use by the next older age set (see Kenyatta, 1971; Launay, forthcoming, reports similar practices in the past in Côte d'Ivoire).

Adolescent ngweko relationships typically involved multiple partners; one avoided seeming selfishly possessive. In Nigeria, Igbo adolescent girls and boys paired off for mild sexual play and "moonlight dancing," sleeping together on mats under the watchful eye of parents or the girl's betrothed husband, with the girl decked out in a special garment tied tightly through the crotch (Ottenberg, 1989). Where full sexual relations were allowed for

young men, as they were for the Masai in Kenya, they were limited to nonmarriageable, uncircumcised prepubertal girls, with an explicit prohibition on premarital pregnancy (Standing and Kisekka, 1989). The sanctions against unacceptable sexual practices that could lead to reproduction among unprepared youth could be extremely severe, including enslavement.

For young men, then, societal determination to separate sexuality from legitimate reproduction created an expectation of male sexual discipline. Youths were expected to acquiesce to long periods of celibacy or to exercise restraints on full penetration. Even after marriage, a man could spend long periods practicing self-control so that his wives could observe up to three years of postpartum abstinence for the physical and spiritual welfare of the child (Caldwell and Caldwell, 1977; Page and Lesthaeghe, 1981). And in areas where women married as young as 11 or 12, marriage to a young bride did not always entail immediate sexual access. In northern Nigeria, where radical female circumcision is practiced, younger men, especially those with no other wives, were viewed apprehensively as prospective husbands for virgins; older men were deemed more likely to manage a careful approach, taking up to 18 months to achieve full consummation (Murray Last, personal communication). Combined with these constraints imposed on certain phases of life, however, were numerous medicines to promote male potency and sexual stamina in other phases of life, as well as potions for women to increase attraction and receptivity (Longmore, 1959; Keller, 1978).

In sum, the expression of sexuality and reproduction were culturally endorsed; but they were supposed to occur only within sanctioned states. Indeed, a theme that emerges repeatedly in ethnographies of the past is the number of socially defined obstacles to reproduction: injunctions to delay the expression of sexuality until after initiation, to avoid reproductive sexual relations with an inappropriate partner, to postpone full sexual exposure for an immature married girl, and to observe postpartum abstinence during breastfeeding. For women the pressures to reproduce while adhering to these strictures resulted in a quick entry into active sexuality and childbearing, followed by long nonreproductive spaces between children. For young men, these strictures appear to have resulted in long periods of nonreproductive adolescence. When nonsanctioned births did occur, it was generally the women and their offspring, rather than the men, who suffered the most serious consequences.

These injunctions defined a trajectory of male sexual life that was very different from the traditional European image of a two-step sequence of celibacy, or near celibacy, followed by routine sexual access to one woman. For African men, the peaks and troughs of sexual access were much sharper and recurred throughout life: Routine sexual access was guaranteed only to high-status older men with many wives. Sexuality was paced and built up

over time during a long process of training and of economic, social, and ritual investment.

How have the dramatic changes in the wider society and economy that are described in Chapter 3 in law, religion, and education—affected not simply marriage but entry into sexual life and childbearing? It is dangerous to draw conclusions about change. African societies of the past have been portrayed romantically, as harmoniously integrated, sharing, and free of major inequities; by extension, the present is often cast as a moral waste-land. Similar dichotomies have applied to descriptions of sexuality. Some sources portray a puritanical past in Africa; others see permissiveness. (For a detailed review, see Caldwell et al., 1989; see also Le Blanc et al., 1991, and the Caldwell et al. response, 1991.) Without question, there were deviations in the past from the ideals just as there are certainly deviations today—quite likely many of them. However we interpret the ethnographic evidence, marriage per se probably was less important for sanctioned reproduction than were acknowledgment of paternity and the parents' ritual preparedness.

Many studies of the present allude to increasing sexual activity and unsanctioned reproduction among unmarried adolescents. Some are based on assertions that experiments in nonreproductive sexuality have declined (Kenyatta, 1971; Launay, forthcoming). Others observe the rise in the phenomenon of ''outside wives'' (deuxième bureau or femmes libres in French) in urban centers in diverse areas: the copper belt, Zambia (Epstein, 1981), Lagos (Karanja, 1987), Nairobi and Dar es Salaam (Southall, 1961; Obbo, 1987), Lusaka (Schuster, 1979), Freetown (Harrell-Bond, 1975), and Accra (Dinan, 1983). The seeming increase in pregnancy rates among schoolgirls, and the resulting disruption of their education, have been another major cause for concern (Southwold, 1973; Schuster, 1979).

The precise causes of these changes are not clear. Some studies cite the Western ideas that eroded the old checks on sexuality when young people encountered new opportunities to achieve status outside the family. These opportunities took the form of migration to cities, conversion to new religions, school attendance, and legal measures against the military activities of age sets (see, for example, Ochollo-Ayao, 1976). Wangui Njau (personal communication) has articulated the widespread view that Christianity in particular produced a paradoxical effect: In banning customs such as polygyny, child marriage, and ngweko, which ran counter to Western norms of propriety, Christianity ironically spawned more promiscuity.

For young men, the large age gap between spouses and the resultant high levels of polygyny still delay marriage until the late twenties or older. Yet certain values within the broader economic and demographic regime of "wealth in people" that emphasize sexuality and reproduction clearly continue to be powerful. Most notably, both the contemporary media and rumor still present male power as entailing an active and varied sexual life with multiple partners. (See Clignet, 1970; Bleek, 1976; Pellow, 1977; Schuster, 1979; Asante-Darko and van der Geest, 1983; Hagan, 1983; Pittin, 1983; Karanja, 1987; Obbo, 1987, Okpewho, 1987; Ba, 1989). The following passages reflect these mixed feelings:

Although access to fully sanctioned sexuality and reproduction is still limited for young men, elders today complain that many of the intricate controls exercised in the past by adults over youthful aspirations to full sexual expression are being dismantled. The observations by David and Voas (1981:658) of the Fulani of Cameroon reflect the general attitude toward male sexuality: "There is no expectation that young men remain celibate before marriage, nor is any value placed upon the fidelity of married men, particularly when away from home." Throughout the region, full premarital sexual relations of men are now taken virtually for granted.

The actual levels and forms of male sexuality are not well documented empirically, except possibly in new research on AIDS. In particular, we lack information on the age at first sexual experience and on the age at first marriage for men. But the combined logics of the ethnography of precolonial demographic regimes, the history of colonial social and economic change, and the demography of current age and marital patterns strongly suggest that, contrary to popular perception, elders' control over youthful fertility

may have changed more for young men than for young women. In the past a young man might have had little opportunity for full sexual expression until his late twenties. Today opportunities for schooling and employment outside the family give young men more independence, and they aspire to a sexual access that the elders' power can no longer deny.

Substantial anecdotal evidence supports the views of society elders that the age of full sexual expression for men has fallen (see, for example, Launay, forthcoming). A Yoruba man complained that the father of his daughter's child was "just a boy," and therefore not able to marry or take on responsibility for the child; and in western Nigeria, members of the older generation charge that an increase in sexual expression is one effect of unemployment for youths—youths with "nothing to do" (Guyer, unpublished field notes).

Church teachings about abstinence have meant that young people are no longer taught the practices of incomplete intercourse that were once permitted (Scheub, 1988), or that they may be unwilling to use them. The life history of a Xhosa woman born at the turn of the century includes a description of courtship with imetsha, incomplete intercourse, starting before puberty (Scheub, 1988), but it is said to be little practiced now. The initiation schools of South African peoples, in which boys learned sexual skills, were already controversial and in decline by the 1950s, and where they existed they were least relevant for urban children. Urban Africans commented that "the excellent custom of ukumetsha has been lost" (Longmore, 1959:46, 157–160).

Changing employment possibilities for young men also may allow greater sexual freedom. Work opportunities typically draw young men away from parental authority, as short-term wage workers, loosely attached apprentices, and self-employed hustlers of many kinds. At the nodes of the transport system, the numerous teenage boys who load lorries, entice passengers, collect fares, and sell anything from mothballs to mousetraps are quite aware that some women offer their sexual services to make a living. With the high value that the society continues to place on sexual expression, these young men are hardly likely to remain aloof.

Relaxing restrictions on sexual expression for young men would substantially intensify the sexual demands on teenage women. And, although men of senior generations continue to seek partners of varied ages, they now compete more directly with young men for younger women. The result of all these trends may be a great increase in the sexual demands men place on younger women, both married and unmarried.

Although the age of full sexuality may have fallen for men, young men nevertheless appear to be avoiding the social responsibilities of paternity before they are considered ready. Court cases from the Beti area of southern Cameroon provide a significant example (Guyer, unpublished field notes).

Although many men could not afford bridewealth during the early 1960s, when the price of cocoa plummeted on world markets, they still fathered children, who were officially registered as père inconnu (father unknown), even when the community knew perfectly well who the father was. As the economy recovered, court cases involving official recognition of children suddenly picked up. Particularly interesting are the claims made for daughters, who were now reaching the age when parents might expect bridewealth on the occasion of their marriages. The temporary decline in marriage and the legal status of pères inconnus created a space in which men could be selective about the recognition and support they provided to their children. Most children from this generation may well have been claimed eventually; certainly the majority were materially supported to some degree. But fathers could now take a discretionary approach to their children through the official vital registration system.

Just as recent changes appear to have given young men more independence in competing for sexual partners, women, too, enjoy new options. Still, married as well as unmarried women face sexual pressures, though they are altered from the past. For married women, an increase in men's sexual demands may impose more risks from pregnancy and childbirth in societies in which the age at marriage is very young. Patterns of initiating sexual relations slowly after marriage to a young bride may be disappearing, resulting in acute health risks for young women.

In The Gambia, a young woman declared that a husband may try to take possession of his bride sooner than in the past by withholding bridewealth until he gains full access, thereby forcing financially hard-pressed parents to turn over their daughter sooner than they would have in the past (Bledsoe, unpublished field notes, 1992). Of course, this strategy may have been used in the past as well. An alternative interpretation might investigate possible declines in older women's authority. In polygynous systems in which marriage was contracted during the wife's childhood or even during her infancy, the loss of older women's authority would leave husbands alone to make decisions about consummation and sexual frequency with very young brides. And in areas in which women's societies once took several years to initiate girls, the new demands of schooling, by cutting down the initiation time, may lessen the control older women have over the timing of young women's marriages.

Although married women may well feel the effects of the reduction in the age of full male sexuality, these effects appear to fall most heavily on unmarried women. They are compounded because the erosion of control over the sexuality of young males coincides with the new opportunities that

young women have to pursue careers outside the domestic sphere—opportunities that make them less eager to rush into marriage and childbearing.

A key theme underlying this chapter has been the observation that the sexuality and fertility of men are as important to understanding childbearing as those of women. Adhering to strictures on sexual activity obviously required the compliance of men. Yet for obvious reasons, fertility studies usually accord women the most analytical attention, leaving men in the shadows. The limited information that we have about young men gives us more perspective on two things: the intensity of sexual demand on teenage women, and the likelihood that a father will recognize and support his children. Although adolescent boys are still considered too young for socially sanctioned reproduction, the more they enter full sexual activity, the more likely will their young partners bear children outside a sanctioned state, whether this consists of marriage, completed initiation, or paternal recognition.

As for the implications of these observations for the handling of premarital births, husbands and kin appear to be increasingly unwilling to acknowledge "illegitimate" children (see, for example, Launay, forthcoming). To understand why, we can profit by looking at the past. The evidence suggests that men were glad, on the whole, to have children, no matter who the actual father was. Ill-timed sexual experiences or illegitimate births could often be rectified or covered up by hasty marriages to young suitors or older polygynists. Today, having a premarital birth makes it difficult for a young woman to find a man willing to marry her, since she is stigmatized as promiscuous; in the past, such women would have easily found older men willing to marry them (Akong'a, 1988). To be sure, ranking and marginalization, processes that intrigued Kopytoff and Miers (1977), were standard modes of operating in most precolonial contexts. Wives, chiefs, slaves, and children—all were ranked relative to each other in terms of power and resource allocation. Yet a child who was born under questionable circumstances, once ritually purified, could usually be incorporated into the family because of the general value people placed on children.

How does the past, as interpreted in these ways, compare to the present, as we know it? Ranking and marginalizing are critical phenomena in modern Africa as well, but they are profoundly underdocumented with respect to their demographic concomitants for women and children. Especially in countries experiencing economic deterioration, families may have to differentiate more sharply the opportunities available to their members. Not only do men appear less inclined to marry multiple wives; they also may be less interested in acknowledging all the children they father, at least in the short

run (see also Guyer, 1984, 1986b). It is unclear whether this decline in willingness to acknowledge social fatherhood stems from efforts to avoid the economic burdens of fatherhood, the problems of urban polygyny, or a substantially earlier age at full sexuality.

Some young women are trying to delay marriage and childbearing, and paternity is growing more costly for young men. Young men are gaining a certain degree of economic independence from gerontocratic control and perhaps even a sense of impunity for their actions. Yet they are not necessarily able, or willing, to support a wife or child. Africa has not developed in its cultural repertoire the idea of the utterly unknown and absent father familiar from the history of European illegitimacy. But in countries experiencing economic decline, it is not difficult to see how paternal support could attenuate to recognition alone, and recognition could attenuate still further into neglect for many offspring of adolescent fathers. Gyepi-Garbrah (1985b) confirms that in Nigeria, many fathers deny responsibility and disappear. (See Iliffe, 1988, for a stark picture of the life that marginal children can face in Africa.)

Having described some of the ways in which reproduction is legitimated through ritual preparation or learning, proven fealty to in-laws, or the acknowledgment of paternity, we turn to a new domain of preparation for adulthood that has received disproportionate research attention in fertility studies: formal schooling. Although past modes of preparation are increasingly being eclipsed to make room in childhood for schooling, Chapter 5 shows that the longstanding cultural emphasis on delaying parenthood to complete preparation for adult life persists with surprising vigor in this new cultural form, which has in effect become a new form of nonreproductive training for adult life.