4
He Who Eats with You
Kinship, Family, and Neighborhood

Mla nawe hafi nawe ila mzaliwa nawe: He^[1] [who] eats with you [will] not die with you unless he was born with you.

This proverb suggests the central place of kinship in Swahili life and a fundamental reason for that centrality. First, there is allusion to the widely shared concern that associates will accept your hospitality (and with it, your friendship) but, in the end, desert you. This is in contrast to kin, who are, the proverb asserts, the only ones you can count on. It is not that kinsmen are understood as always trustworthy and loyal but only that in a difficult world, they are the ones who are most likely to stay with you.

Ndugu muwi afadhali naye: [A] bad kinsman [it is] better [to be] with him [than to be without him].

No matter what you may think about a kinsman, he or she will always be your kinsman, and, even if he is bad, as the proverb says, it is better to have him than to be without him. On emotional, material, and social grounds, kin should not be denied.

Damu isipowasha hunyeza: [If] blood does not burn, [it] itches.

This proverb emphasizes the undeniability of kin ties. The proverb is heard about people who have reservations concerning a kinsman or do not seem to others to be very close to him or her. Sometimes the proverb is used to warn someone who is thought likely to say something unfavorable about a person in front of another who is that person's kin. Even agreeing with someone's criticism of his own relative can prompt this warning, because it is generally

believed that whatever reservations one may have about a kinsman, the tie with him or her will always remain in force. This applies to all kin but is heard most often concerning one's own and one's parents' siblings.

In this community the emotional attachment to kin is very rarely severed completely, but even if it were, there remains an unavoidable social attachment. This is based in the fact that closely related people, especially nuclear family members, are categorized into a common status as "members of a family" (ndugu or ukoo mmjoa ) and, as the next proverb indicates, have a common social and material fate.

Mchuma janga hula na wakwao: He [who] earns calamity, eats [it] with the people of their house [i.e., his family].

Kin are by no means viewed as always good to one another, but the closeness of their interests is understood, as the following proverb shows rather trenchantly, to lead them to work together and to avoid doing anything really harmful to one another.

Meno ya mbwa hayaumani: [The] teeth of [the] dog do not bite [harm] each other.

This proverb makes a particularly important point, since, as several informants who discussed it with me agree, the strong emotional and social bonds and the commonality of interests that unite kin also puts them in a position to harm one another as no outsider can. The proverb does not mean that kin never harm one another but rather that although people may turn their hostility, acquisitiveness, or indifference on others, even the worst of them (i.e., dogs, which most Swahili, in common with many other Muslims, find revolting as well as sources of ritual pollution, najisi ) find it difficult to direct at their kin. Like a set of teeth, kin work as a unit and do not usually harm one another, whatever they may do to others. Because of their general and pervasive importance to one another, however, kin can be more bothersome than anyone else can be.

Chawa akuumao mbwa nguni mwako: [The] louse that bites you is inside your clothing.

Taken together, these proverbs suggest the broad and pervasive social, material, and emotional ties among kin and their enduring quality. It is important to note that the proverbs are all general indications of the importance of kin relations and the unity of interests among kin. They do little to indicate what kin can be expected to do or not do. In fact, insofar as they provide

any guidelines for behavior at all, they are very broad guidelines that counsel remembering kinship and the permanence of its ties without reference to what is actually to be done by whom for whom in specific circumstances. This is not only true for the proverbs I have quoted here but for all the others concerning kin that I have encountered (e.g., Scheven 1981:325, 328, 329).

As the discussion of Swahili social life proceeds in subsequent chapters, it will become clear that the participants in wide-scope and long-enduring relationships such as those involving kin and neighbors share broad and general expectations regarding each others' behavior more than they do specific ones. The most generally shared specific cultural elements in these relationships are the "identifiers" (see chap. 1), but my main concern here is less with specific understandings than with the general ones that are the main contents of the relationships between those identified as belonging to family statuses.

Kin, Household, and Nuclear Family

The importance individuals give to kin relationships generally is intensified as the kin ties are closer. It is notable that a number of the proverbs affirming the importance of kinship invoke eating and commensality in ways informants with whom I discussed the proverbs agree show closeness of relationship.

It is the household group, rather than some more inclusive collection of kin, that regularly eats together (although in many homes, the commonly prepared food is eaten in two seatings: first the men and then the women). And it is the household group that contains the kin who are the most broadly important members of the generally important category, kin. Kin relations are said by everyone with whom I talked to be more important in most activities than any others. Of these, the ones involving fellow members of the same household are more important yet.

As we will see below, the household is the material, social, and emotional base for much of what its members do, and more than three-quarters of Swahili households include or are solely composed of the nuclear family (see table 2, below). In examining cultural sharing among household members (chap. 5), I limited the survey to households where the person "in charge" (mwenye amri ) was a member of a nuclear family including both spouses and at least one child considered by them (and by the child) as their biological offspring. This was an attempt to increase comparability.

This focus on the nuclear family is justified by the prevalence of this group among the Swahili as well as by the interest in examining in a natural setting a grouping in which the sharing of understandings about the group, its members, and their relationships would be expected to be in the upper part of the range of sharing.

Narrowly defined groups focused on technical tasks (e.g., a surgical team) might be expected to have members who share more completely the understandings about their work and each member's role in that work. Monasteries or other religious communities and other highly specialized groups may have an extremely high level of sharing among members in some respects, but in dealing with others concerning the broad array of matters that make up ordinary everyday life, the nuclear family would seem to call for at least as high a level of sharing as would be found in any other social group and to provide an ideal situation for mutual socialization as a basis for that sharing.

In chapter 5, we will see that the elements of family culture, those understandings having to do with the nuclear family and how its members do and should treat one another and with what the group should and does do, are only partially shared, even by members of the same nuclear family. One reason for this could be that the nuclear family is of limited scope and importance in Swahili life, but, as we will see, this explanation is not characteristic of the nuclear family in this society.

If group interests and the relations among members were restricted in scope and intensity as those, say, of the residents of a boarding house are, even members of families that continue functioning for long periods of time (and only such families were included in the survey work that formed the basis of the study of sharing) might share only a limited range of understandings about the group and its constituent relationships. Since, however, the interests of the Swahili nuclear family are broad and of intense concern to members and since family relationships are seen as vitally important, limits in the sharing of understandings about these interests and relationships cannot be explained as based in substantially restricted joint concerns and involvements.

What is said about the importance of the nuclear family—and we will consider this more specifically in a moment—should not be taken to suggest that all households are composed of, or even contain, a nuclear family. Fourteen of the 111 households surveyed for the census I made did not include both spouses. Most of these fourteen households are headed by divorced or widowed women, but, despite the emphasis given male authority by the Swahili and the support this has in Islam, these households function in many ways much as do those with a complete nuclear family. They are evaluated by outside community members, according to the evidence I have, on the basis of the same broad understandings that apply to households whose core includes both spouses.

Similarly, roughly half of all households—whether the households include a full nuclear family or not—include one, rarely two, adult, nonnuclear family kin. These "extended" households, too, give every appearance of operating according to the same broad understandings that apply to the other half of the population of households whose only adults are the spouse-parents.

It is not surprising that there is variation in household composition and

that a substantial minority lives in a household without a full nuclear family present. Death and divorce sunder nuclear families in Old Town as they do elsewhere. But this does not diminish the importance of the nuclear family even for those whose household does not contain one. None of the many individuals I talked to who lived in nonnuclear family households thought that the composition of their household was an ideal one. Most viewed their situation as more difficult economically and socially than it would be if there were a full nuclear family in their household, but the truncated (usually because of divorce or husband/father's death) household was their main source of emotional support, material assistance, and social relations. In this they were no different from the majority who live with their spouse and children or parents and siblings.

Whether a nuclear family or not and whether including "outside" kin or not, the household members are much involved in each others' activities and interests. The mother often is, as in many societies (e.g., Bott 1971:69 ff., Young and Willmott 1973:101–102), the hub of activity, and everything her children do is of immediate and central concern to her. At the same time, her views and assistance are usually crucial to the children, at least until they have established their own nuclear families and, most often, even after that.

The mother/wife is also often much involved in her husband's activities. As we will see in chapter 10, men depend on their wives as they do no one else, and if wives have a considerable set of interests (many of them shared with the children) separate from the husband, he is still the most important adult male in the vast majority of households. In many families, men have rather distant relations with their children, especially with the sons, and deal with them mainly through their wives. However, the emotional and material ties between fathers and children are deeply felt and highly influential. As we will see, there is a substantial and sharp division of labor and leisure in the Swahili nuclear family, but that group is of the first importance to all its members despite the differences according to their statuses in how much and in what ways they participate.

In spite of this, the data examined in chapter 5 show that there is only limited sharing of the cultural items directly concerning life in the nuclear family household among the members of that very same group. Among community members belonging to different, but equally stable, families, the sharing of cultural elements concerning the nuclear family is even more limited. These findings are taken as the basis for the view, quite fundamental to this whole study, that extent of cultural sharing is only one, and often not the most prominent, of the contributors to the effectiveness of the groups in dealing with their members' needs and interests.

In fact, other studies of cultural sharing among Swahili not belonging to the same nuclear family (to be seen in chaps. 7 and 9) confirm the limited nature of cultural sharing in this society. Work by others suggests that the

same is true in social groups in other societies (e.g., Pelto and Pelto 1975; Holland 1987a ; Kessing 1982).

To see the nuclear family's place in Swahili society and to understand more of the nature of that society, it will be useful to consider the nuclear family in the context of kin relations in general.

Swahili Kinship

As table 1 shows, Swahili kin terms do not distinguish cross from parallel cousins, but they do have different terms for matrilateral as opposed to patrilateral parents' siblings. As noted earlier, the Mombasa Swahili have no unilineal descent groups despite the use of the term "clan" by English-speaking group members. In fact, the only groups of this sort with members living in Old Town are those made up of Arab immigrants who are either fairly recent arrivals from the Persian Gulf or members of the Mazrui group, which has been in East Africa for a very long time but retains its clan, sensu stricto , organization.

The patrilineal mbari is often referred to as a "clan" in the literature (e.g., Pouwells 1987:79). Mbari are not localized (and may never have been), do not and never did own property, are not exogamous, and, in short, have none of the attributes of a clan other than a belief in relationship through fathers (not always excluding fathers' sisters' offspring). For the Mombasa Swahili, in fact, there are no corporate unilineal descent groups.

As far as contemporary relevance of the mbari is concerned, there is no activity of any kind attributed to them in the last four decades or more prior to which they are said to have played a minor role in wedding rituals.^[2] I could find only a few elderly people who knew which of the mbari they belonged to, and a number of younger informants said they were not even familiar with the word "mbari."

The extended family is commonly referred to as mlango (lit., door), but ukoo is also used even though it refers to the nuclear family in some contexts. The mlango is basically an ego-centered category of bilaterally related kin whose members do little or nothing requiring organization or general participation. Ad hoc groupings, usually made up of women kin and neighbors, form for particular, limited purposes (mainly to prepare for and participate in weddings, funerals, the now very rarely publicly celebrated circumcisions [tohara ], and maulidi, or public readings of the life of the Prophet followed by refreshments). These activities aside, group activities limited to or arranged by kin drawn from beyond the nuclear family are nonexistent. Participation in such jointly arranged activity as is carried out depends as much on the affectional ties and current interests of members as it does on their kin statuses, and unrelated neighbors are often as prominent in them as kin are.

Save for living parents and their dependent children, who share the proceeds from wealth inherited by one of the parents, property ownership is virtually never shared among kin, even siblings. When a father or mother leaves land, a house, or an apartment building to his or her children, it is either sold and the money divided according to Islamic inheritance laws or the building or land is physically partitioned so that each heir has his or her share as separate, personal property.

People of means sometimes create trusts (wakf ) supervised according to Islamic law to benefit their children and grandchildren. Several informants said that this was done to conserve the property (only its income is available to the heirs) and, also, to lessen the chances that an adult brother (or, rarely, some other surviving kin) of the deceased would misappropriate the inheritance while serving as the children's guardian until they reached their majority.

Disputes about inheritance are settled by local kadhis (judges of Islamic courts) according to Koranic principles. A Swahili language book presenting the principles of the Islamic laws (sheria ) of inheritance is taken as the final arbiter of disputes and is consulted both by kadhis and by lay persons concerned with questions of inheritance (Kasim el Mazrui 1952).

Parents and Children

There are some generalities regarding relations among kin^[3] that can be inferred from informants' statements and observed behavior. These include respect and deference for parental generation kin which diminishes somewhat as the ego generation kin reach adulthood but does not disappear until and unless the parental generation kin becomes senile (pishwa ), and even then some indications of respect and deference are retained. Outside the nuclear family, there is general similarity in behavior toward parental generation kin regardless of sex, although, given the pervasive division of the sexes and the isolation of women (tawa ), there is generally more interaction within one's own sex.

Relations between mothers and children are usually closer, freer of conflict, and less restricted than relations between fathers and children. Fathers are said—including, sometimes, by themselves—more often to get along well with daughters than with sons. In fact, boys and young men rather often make use of their sisters' good relations with their father to get the father to do things the sons want him to do. A daughter asking her father to do something is likelier than a son, several informants say, to succeed even if what she wants is for her brother.

Some young men report excellent relations with their fathers, and observation in these cases is in accord with their reports. Others, however, say that although the father-son relationship is warm and close when the boy

Table 1.
Kin Terms

Table 1.
(continued)

is small, as adolescence approaches, the relationship becomes more distant and/or characterized by conflict.

I have talked to a number of adolescent boys and young men who have critical, even bitter, things to say about their fathers. A number of sons criticized their fathers for being too strict, too old-fashioned, and too unwilling to provide needed funds. I have never heard a girl or young woman criticize her father. However, several have told me that their parents (wazee ) are too strict or old-fashioned but without being willing (or able?) to differentiate between their fathers and their mothers in this respect. Aside from whatever share they have of the general criticism of parents, mothers are not, in my experience, subject to open criticism by either daughters or sons.

A few boys and young men leave their homes before they marry, and the reasons given are uniformly based on difficulties in getting along with their father or, somewhat more commonly, their mother's husband whom she married either following divorce or the biological father's death. In all seven of the cases of this sort that I was able to record, the son visited his mother even though, as in two instances, the husband/father told her not to see him.

As will be seen below, mothers almost always have greater involvement than fathers in their children's lives. In some families, the fathers are mainly peripheral to the children's activities except for providing (or refusing) money for these activities, including schooling.

Relations between parents and children are by no means free of conflict. Mothers' day-to-day relations with their daughters are fraught with conflict, and much of the most colorful obscenity in the Swahili language (and it is a language rich in abusive resources) is directed by mothers at their daughters (Swartz 1990a , 1990b ). I have no evidence to indicate that fathers and sons or fathers and daughters exchange insults, but tensions in these relations—often related to differences regarding the use of money—frequently are expressed in silences and withdrawal.

Some conflict is taken as natural. A proverb says Pesa zikiwa mfukoni haziwati kugongana (Money that is in [a] pocket does not stop knocking together). This emphasizes the inevitable and constant minor conflict among those who are close to one another (in the same pocket) and the harmlessness of that conflict. Another proverb asserts that the familiarity that allows minor conflict is strengthened by that conflict: "Nyoko, nyoko—ni faida ya kuonana " (Mother, mother [this is an archaic form that appears only as part of fairly mild obscene insults and refers to that insult here]—is [to the] advantage of [a] relationship).

But the recognition that some strife is natural and harmless in parent-child (and other household) relationships does not obviate the fact that serious ruptures can and do occur. As noted, there are sons who leave the parental home before marriage and not in order to work at a distant job. I know of no Swahili woman who has left her parents' home as dissatisfied sons sometimes do, but

the restrictions on women are such that departures are extremely difficult unless the woman marries or goes to the house of a kinswoman, in which case I would not be likely to hear that family trouble was at the root of the move.

As understood among the Swahili, it is a tenet of Islam that a child cannot go to heaven if the parent does not have radhi for the child when the parent dies. "Radhi" means a blessing and comes, informants report, from the parents' satisfaction with the child's behavior. A parent cannot control radhi; it is a natural consequence of the child's behavior, and a child who has behaved intolerably toward a parent will not get radhi regardless of how great the parent's love for him or her may be. Radhi need not be given explicitly, and at death a parent may give or withhold it without necessarily knowing that it has been given or withheld.

Since the Swahili, including the members of the younger generation, are almost all deeply religious Muslims, the importance of radhi would seem a significant resource for parents in their relations with their children. To some extent this is true, and an element of children's general wish to please their parents is, they say explicitly, based in concern about this blessing.^[4] Much more active in shaping relations with parents, however, are the very strong bonds that unite parents and children. A mother is said to feel deep sympathy and love for her children because of kitei (Johnson 1959 [1939] renders this as kite ), the pain she suffers at the birth of her child. Sometimes a person who refuses a child's request or withholds sympathy from a child is told, "Hukuzaa " (You have not given birth). Even fathers are said to have given birth and to have sympathy and love for a child because of this.^[5]

Although fathers are said genuinely to love their children, some part of this is understood—at least by some male informants—as a consequence of their love for the child's mother, their wife. A proverb used mainly with respect to stepchildren but explicitly said to apply to own children as well maintains that a man who loves a woman must also love her children: Mtu akipenda koa, hupenda na kilicho ndani ([A] person [who], if [he] loves [the] oyster shell [customarily or usually] loves that which is inside [it]).

Mother's Siblings and Father's Siblings

The mother's brother, mjomba , occupies a fairly distinctive position in the parental generation, and informants agree that this man is more likely to be sympathetic and helpful than a father's brother, Baba ya pili [lit., second father] or Baba mdogo [lit., little father]. Several informants say that if one's father dies while one is young, it is better to be in the charge of a mother's brother than a father's brother because, given the partrilateral emphasis in Islamic inheritance laws, the former is not in a position to try to appropriate your share of your dead father's estate as the latter is. Further, and at least

as important, there is more likely to be "love" (upendo ) on the mother's side of the family. The mother's brother's wife is referred to by a distinctive term, mkaza mjombe .

Like mother's brother, father's sister is referred to by a term, shangazi , that distinguishes her from other kin of the parental generation, although there is no distinctive term for her husband. However, no particular distinctive qualities in the relations between people of either sex and their father's sisters were observed or reported by informants. If anything, it is my impression that mother's sisters play an active role in people's lives more often than father's sisters do. When I mentioned this impression to informants, they uniformly agreed and said that it was due to the fact that when children are small, they go with their mothers when the latter go visiting. Since mothers visit their sisters more often than their husbands' sisters, the children become more accustomed (zoea ) to them and, therefore, have closer relations with them.

Siblings and Cousins

Own generation kin are treated quite differently according to sex. There is often a good deal of interaction among same sex, same generation kin, although it is my impression that sisters, sisters' daughters, and similarly related women are more likely to spend a great deal of time together than comparably related men.

Still, in many families, brothers are together a good deal before they marry. The elder serves as a mentor for the younger, and the younger helps the elder in whatever tasks the latter may be involved in. With marriage and the beginning of their own families, relations become more formal and distant, and sometimes competitiveness, even mutual hostility, emerges.

Relations between sisters are frequently close and warm in childhood when they are almost constantly together as a consequence of the isolation of women, which led—and, to a slightly lesser extent, still leads—to girls and young women spending much of their time in their own house and those of female relatives and neighbors whom they visit with their mother. After marriage, relations between sisters generally remain quite positive and lack the competitiveness that sometimes is seen between brothers.

Brothers and sisters are together less as children than are same sex siblings, since while the girls stay in the home when they are not in school, the boys, after the age of seven or eight, are either in school or wandering the neighborhood and playing games with other boys. There is some conflict between brothers and sisters deriving from brothers attempting to control their sisters, especially with regard to the sisters being allowed to go out of the house. Sisters often depend on their brothers to do errands for them which involve going outside, and brothers depend on their sisters for part of the cooking, cleaning, and mending they require. As might be expected, this

work for brothers by sisters sometimes leads to resentment, but it also serves as a check on brothers' attempts to control their sisters' activities.

Matrilateral Association and Affection

Several informants have said that both men and women are closer to their mother's kin than to their father's as a consequence of closer emotional ties to the mother and of going with her to visit her kin when they were young. My observations (as well as the data in the survey reported in chap. 5) tend to bear this out. Nevertheless, beyond the confines of the nuclear family, common interests and personal affections are the most powerful determinants of which kin associate often and closely with one another and which have more distant relations.

In accord with this, group members consistently affirm that there is no general difference in one's relations with father's brother's children, mother's brother's children, father's sister's children, and mother's sister's children. All own generation kin can be referred to as ndugu (the same term, in fact, is used for all kin regardless of generation or nature of relationship), but there are terms that distinguish between brothers (kaka and mdogo are fairly common) and sisters (dada or the more Arabized ukhti ).

There are limited terminological differences found in alternate terms which distinguish the children of one sort of parents' sibling from another and from own siblings, but there is little basis for believing that these differences have much sociological or psychological significance. They seem mostly to be used to clarify a relationship for the benefit of visitors or anthropologists. The alternate terms are used only in reference and clarification and mean, literally, "child of mother's brother," "child of father's brother," and so on.

Cousin marriage is quite common, and more than a third of all the marriages about which I could collect genealogical information were between kin.^[6] I have been told that it does not matter if the cousin you marry is related to you through your father or through your mother (although a number of informants are quite emphatic about how important the difference is to "Arabs"), but that it is slightly better to marry someone who is not the child of an actual sibling of either parent. In fact, two marriages about which I have relatively full data involved the child of a parent's true sibling, and neither of these caused any observable comment.

Generational Kin Term Uses

The terms for members of the child generation are basically the same for all kin on both sides. The females are referred to as binti and the males as mwana, with the latter term sometimes used for children of both sexes. I

found no use of terms differentiating among child generation kin on any basis other than sex, but explanations are added ("the child of my sister" or "the child of my mother's brother's son") when needed.

The terms for parents are extended in an interesting way to members of the child generation. A crying child is comforted by, among other ways, being addressed as baba (father) or mama (mother) by his or her parents or older siblings. Informants say this is done to give the child respect or prestige (heshima ) and, thereby, cheer him or her. The fact that little boys of ten or less are sometimes addressed as Bwana Mkubwa (or the diminutive of that, Chubwa, meaning Great Sir or Great Lord) seems to serve the same function, since this same term is used not only for rich men, political officials, and important employers but also for the father's father, the mother's father, and, sometimes, for others called Babu (grandfather).

Child terms are extended outside the family to children, particularly members of other ethnic groups, from whom favors are being asked. Swahili women, especially older ones, have a particular, ingratiating inflection they use in addressing an unrelated child as Mwanangu (my son) when they are trying to get the child to run errands for them.

Grandchildren are spoken to in a quite familiar way, and relations between grandparents and grandchildren are ideally and actually warm, friendly, and as nearly egalitarian as they can easily be given the differences in resources and physical abilities.^[7] A grandparent who is present when a grandchild is scolded or punished will, especially if the punishing parent is the grandparent's child, take the child's side and urge the parent to forgive him or her. I heard a grandmother tell her daughter in what appeared a serious way that it would be better if she, the grandmother, were punished rather than the grandchild for what the latter had done to annoy the mother.

Kin and Expectations

Within the substantial range of variation in relationships involving kin, one thing is universally true: the main expectations in these relationships are general rather than specific. These general expectations are the heart of relationships not only within the household and nuclear family but also with fairly distant kin such as grandparents and grandchildren, parents' siblings and their children, and, to a considerable extent, neighbors who have lived near the family for long periods.

There are also specific expectations in these relationships, of course. For example, wives and mothers and, to a slightly lesser extent, sisters are expected to prepare meals for the household and to see to such domestic tasks as laundry and cleaning.

Husbands and fathers are expected to provide money to the rest of the

household and, together with sons and brothers, to do most shopping and errands involving any substantial traveling around the town. These last expectations involve highly specific activities such as producing cooked food at a particular time or paying a particular bill. There is usually little or no difficulty for the participants in a relationship in determining whether or not specific expectations have or have not been met, although there may be difficulty in agreeing on the significance of meeting or not meeting them.

The Importance of General Expectations

The main expectations in these relationships, the ones that participants most often use as the basis for evaluating performance, are general rather than specific. Establishing whether general expectations have been met or not is much more difficult than doing the same for specific expectations and, at the same time, is often more important. A person who is judged to meet the general expectations in his or her kin status will be positively evaluated in the status by those in relationships with him or her even if that person is not particularly adept at meeting some of the specific expectations in the status. The reverse is less often true.

Since the difference between general and specific expectations is one of degree rather than kind, it is not surprising that the two shade into one another. Specific expectations such as the wife's housekeeping or the husband's provision of money become most important when they are seen as affecting general expectations. A key difference is that general expectations can be met in a variety of ways, which is not so true of specific expectations. A husband who loses his job may fail to meet the money-providing expectations but still be evaluated positively in the spouse role if the wife recognizes other bases for viewing him as loving, conscientious, and having good intentions (nia, see below). The same is true for a wife who cooks badly or fails to keep the house as the husband might want it to be.

General expectations may include telling a child or, sometimes, a spouse about mistakes they have made and how to correct them, but they are characteristically positive in nature. As will become clear below, "upendo," which can be glossed as "love," is one of the main aspects of relations among nuclear family members and, beyond that, with kin and neighbors generally. Some of the more obvious manifestations of upendo are displaying concern, trust, affection, and a regard for the interests of the object of the love. "Accustomedness" and, less, "one character" function much as "love" does, but for the present discussion, love alone will be considered and the others left for later examination.

It is not always obvious whether particular acts meet general expectations,

and even those directly involved can sometimes make a judgment only by considering a whole pattern of behavior rather than specific acts. In part, this is so because the connection between acts and the meeting of general expectations is so dependent on interpretation that a given behavior and its opposite can be taken to have the same significance. Thus, as we will see in chapter 10, a man shows his love for his wife by giving her money for clothing and such even if he thinks buying the clothing is wasteful and unnecessary. Showing love for a son, however, sometimes involves not giving him money when the same man, now in the status father, thinks the use of the money would be wasteful and unnecessary.^[8]

Nor are opposite behaviors given the same interpretation only when they occur in different relationships involving different statuses. A woman shows her love for her sister (i.e., meets general expectations in the relationship with her) by criticizing her even to the extent that the sister weeps, but love can also be shown by denying to the sister the truth of criticism, perhaps leveled by others, of her. In short, general expectations are illusive as concerns their behavioral reference, and this is not just an observer's problem. Participants in relationships often find it difficult to assess whether or not general expectations are being met, but the assessments are nevertheless made and remade with the future of the relationship, as well as the judgment of the individual, depending on them.

Community members are explicit about the difficulty in assessing behavior in relationships where general expectations are the crucial ones. A number of people have told me that in dealings with those who are close, that is, with spouses, children, close kin, and neighbors, nia can be more important than what the person does. "Nia" can be glossed as intentions, purposes, or general orientation. A person whose nia is "good" (nia nzuri ) can do things that may seem bad but that are recognized, perhaps on reflection, as at least well intended. Understanding whether a person in a close relationship has met the general expectations in that relationship, then, depends not only on his or her behavior but also on the view the other participant(s) in the relationship have of his or her nia.

Just what is called for by general expectations is not always clear until a situation arises involving them. Thus, for example, several young Swahili men were arrested for loitering one evening. It was difficult to tell why they had been arrested since they appeared to be doing nothing unusual but only sitting on a low retaining wall near their houses chatting as they generally did in the evening. Because of the vague reason for the arrest and because they were handled rather roughly by the police without any obvious reason (I saw the arrest), the event caused severe anxiety in the young men's households. On hearing what had happened, one of the fathers rushed from his house to that of a neighbor who was friendly with a number of the city's more important officials. In a remarkably short time, the neighbor left his house,

although he had been about to go to bed, and went to the police station where, after long discussions with police officials, he was able to win the young men's release.

No serious question seemed to occur to the fathers of the young men about the willingness of their neighbor to help them. Nor did the neighbor, then or later, indicate that he thought he had done anything particularly surprising or unusual even though he said he had not actually intervened with the police before on behalf of neighbors or, for that matter, anyone. "We help each other," he explained to me in discussing the incident and his role in it.

The neighbor's intervention is in accord with the general expectation of mutual help and support that is characteristic of relations between kin and neighbors. Thus, when the child of a household is being married, the mother's women kin and neighbors speak of "our wedding" and work impressively long hours helping to prepare for it. When the son of a household is ready for a job, not just his father and older brothers but many of his parents' brothers and parents' sisters' husbands, older cousins, and neighbors will all make efforts—some more vigorous than others, of course—to find him a suitable position.

Precisely what neighbors and kin do for one another is mostly not specified, but the expectations in their relationships are, nevertheless, powerful. Just as a politician is supported because he is understood to provide vague but vital goals and states such as "prosperity" or "peace," so kin and neighbors are positively evaluated when they provide "help" and "support." The Arab owner of a "ration shop" is expected to return goods and change to a customer who says what is wanted and gives him money, but expectations of such a clear and specific kind are not the main characteristic of relationships between kin and neighbors. In those relationships, more accurately, in the statuses of the individuals involved in them, the key expectations are broad and general, giving those relationships a flexibility and an ambiguity that relationships based on statuses involving mainly specific expectations cannot have.

Mutual Choice in Forming a Social "Pool"

In general, my observations and discussions with group members support the conclusion that each Swahili is part of a pool of individuals whose members choose one another for relationships of varying intensity, frequency of association, and duration. Some of these relationships are rather distant, and others involve regular mutual assistance, exchanges of some kinds of confidential information, and strong emotional bonds. Ties of kinship make up a substantial part of the basis for forming this ego-centered pool, but neigh-

bors from adjoining houses are another important element, especially for women.

For men, membership in the pool is also affected by the mosque they regularly attend with a man's closer relations—both with kin and nonkin—often mainly with others who attend the same mosque. Since which mosque a man regularly attends is commonly more influenced by proximity than anything else, neighborhood (in the sense of the general area, about which see below) is important for men but is generally a wider area than it is for women.

Individuals choose and are chose from their pool according to personal preference, leaving the other pool members as more or less available sources of material help, social support, and, for women, emotional warmth when limited or exceptional circumstances make these desirable. This availability is maintained through relationships kept active in occasional visits to one another's homes and joint participation in weddings, funerals, and maulidi.

Particular kinsmen may or may not be active in preparing for one another's life crises or religious rituals, but they will make some contribution and will almost always be present. Men who attend them use their pools as the basis for choosing a baraza. Some men do not attend a baraza regularly, but there is an explicit value on men associating with one another rather than being isolated in their homes, and a main venue for such association, second only to the mosque before and after prayers, is the baraza.

In my limited experience with them, women are more often in close contact with more of their kin than men are. Women seem to draw on the "pool" of their husbands' kin, however, rather more selectively depending on their and their affines' preferences, on the women's relations with their husbands, and on their husbands' relations with the members of his extended family.

Men are generally more distant and more selective in all their relationships save, often, with their wives (more of this in chap. 10). There are men who see a great deal of several of their father's brothers' sons and little of their mother's brothers' sons even though informants report that warmer and more satisfactory relations are likelier on the mother's side. There are mature men, including some with children of their own, who associate closely with their fathers and are involved with them in a range of financial, religious, and social matters. But there are others who see their fathers only rarely and have little to do with them. The same can be said of relations between brothers.

Neighbors

For both men and women, the pool of possible close and/or especially significant associates is not limited to kin. Neighbors, including those who are unrelated, are often among those with whom there are the closest relationships, often involving mutual assistance and strong emotional ties. There is a proverb that, if it lacks the trenchancy of many others, nevertheless charac-

terizes the nature of relations with neighbors as compared to kin: Hallah, hallah jirani kuliko ndugu mli kule (God, God [much used to give emphasis] the neighbor [is more] than [a] kin [who] is over there [i.e., at a distance]).

As with kinship, ties based on proximity are a likely basis for close relationships without there being any necessity for such relationships to develop. To be jirani (neighbors; wamtaa is a less common alternate term) is to have the basis for a close relationship, but normatively, informants agree it need only be a cordial one. Some neighbors are likely also to be kin, and the dual tie increases the likelihood of frequent and close association. Other neighbors, however, may not even be Swahili, and close relations even with such outsiders are, within limits to be described, fairly common.

Old Town is no longer a purely Swahili area and has not been since the railroad to Nairobi and Uganda was completed before World War I. The steady increase in the presence of non-Swahili in the area has resulted in greater and greater population density, with much of the additional housing, as well as some formerly Swahili housing, being occupied by outsiders. The limited census I was able to make indicates that outsiders, mainly Indians and Arabs of various sorts (but most commonly those from the Hadhramot area of Yemen), now occupy more than half of the houses in the area. The eastern part of Mombasa island, which was almost purely Swahili prior to World War I (when the rest of the island was used for agriculture), is now a cosmopolitan area only slightly less polyglot than the area adjacent to Old Town that has become the business district and, to the west of that, a mainly non-Swahili residential and industrial area leading to the modern port.^[9]

Despite its greatly increased non-Swahili population, Old Town is not completely heterogeneous. There are almost no resident Christians there, and almost all of the minority of non-Muslims are Hindus. Among the Muslim majority in Old Town, the Swahilis' fellow Sunni are the largest group, although there are some Shia (all or almost all from India), a small number of unassimilated Omanis of the Ibathi canon, and, again mainly Indian Muslims from other non-Sunni groups. Ethnic heterogeneity is, as this suggests, rather greater than religious.

Relations with Non-Swahili Neighbors

Many Swahili women and some men have close relations with unrelated neighbors, some of whom are from other ethnic groups. This close association with outsiders, however, is almost exclusively with those who are Sunni Muslims. Close relations with Indian neighbors, even if the Indians are Sunni, are not very common unless they are Baluchis (see chap. 2), whom the Swahili view as far more like themselves than members of any other outside group. Close relations with Hadhrami Arab neighbors (and the Mombasa Hadhrami are all Sunni) are, however, by no means unusual.

Close relationships between whole Swahili households and unrelated neighboring households are quite common, and people often speak of their neighbors as being "ndugu" even when, on inquiry, it appears there is no known blood relationship. Such relations are very positively evaluated by everyone I talked to, and the issue of whether "real" kin might be unhappy about close relations between their relatives and unconnected neighbors was dismissed as farfetched. A number of informants pointed out to me that Islam enjoins neighbors to care for and love one another and quoted Prophet Mohammed as saying that the only thing that remained to be done to make neighbors the same as kin was to find a way for neighbors to inherit from one another.

Warm and mutually supporting relationships are extremely common between Swahili families that have lived in the same nearby houses for generations, but even neighbors who have lived near one another for only a matter of years often have wide-scope and intensive relations. Such relations sometimes develop even when the neighbors are Sunni from some non-Swahili group.

In many cases, of course, the members of neighboring households differ in the closeness of their ties with one another. Most frequently, it is the wife-mother in the Swahili household who is close to her counterpart in the neighboring household, and other relations between members of the two households vary from cordial or polite to truly friendly. Occasionally, the relationship between households focuses on the husbands-fathers, but this is far less common and these relations, like most men's relations, are usually restrained and polite rather than close, even if the men sometimes do important practical favors for one another.

"One Character,"
"Accustomedness," and "Love":
The Emotional States Understood as Usual in Close Relationships

Long presence in the same nearby houses is important, but more important yet is whether someone in one household sees someone in the other (always of the same sex and almost always women) as having tabia moja (lit., one character or disposition, i.e., being sympathetic to one another and/or similar in personality). This mutual attraction or similarity is said by several informants to be the surest source of close relations between individuals who are not kin and neighbors. The relatively few men and women who have lasting friendships with people who are neither kin nor neighbors also attribute the relationship to tabia moja.

Tabia moja is far less important for kin. The proverbs at the beginning of this chapter make clear the same thing that one observes and hears from infor-

mants: kin relations benefit from personal compatibility, but if the relationships are within the nuclear family, such personal preferences matter far less. Mazoezi (accustomedness) comes from long, close association, and I have been told repeatedly that with mazoezi, being mutually sympathic (i.e., tabia moja) is not an issue. Tabia moja refers to mutual affection and similarity of interests, while accustomedness makes this unnecessary since becoming "accustomed" to someone renders whatever lack of affection or incompatibility of interests there may be manageable and, eventually, unimportant.^[10]

As discussed earlier, one has love (upenzi ) for the parents and siblings and spouse with whom one lives, or, failing that, especially if they are stepparents or children, one is "accustomed" to them. Selections from the pool of possible associates made up of more distant kin and neighbors, however, is on the basis of personal compatibility, tabia moja. Accustomedness and/or love seems the basis of the household group where constant and close association takes place. Compatibility and mutual affection are enough for the less constantly demanding relations between more distant kin, neighbors, and friends. Fellow household members are rarely characterized as having tabia moja with the speaker, but close associates from outside the household frequently are.

Inclusive Neighborhoods: Old Town Sections

As has been true for centuries, the area of the town in which a family lives (mtaa, pl. mitaa) provides a basis for organizing social relationships. The importance of these now is less than it was twenty years ago, but everyone is actively aware of what his neighborhood is, and it still affects social relationships. Each of these areas (see maps 1 and 2) occupies a fairly substantial territory, far more than the few adjoining houses whose occupancy offers the possibility of people having a close relationship as "neighbors" in the sense just discussed.

The residents of each of these areas are likelier to associate substantially more with those from the same area than with people from other areas, and close ties are likelier within than across area boundaries. Each of the areas is understood as having economic, social, and personality traits that distinguish it from each of the others, with the understandings about each area differing, mainly in emphasis and valence, according to whether they are expressed by the residents of the areas in question or by the residents of the other areas.

As noted in chapter 3, each of these areas was a part of one or the other of the two sections that were the basis of community organization until a few decades ago. This organization has lost its effectiveness as a basis for joint

Map 1.
Mombasa showing Old Town North.

Map 2.
Mombasa showing Old Town South.

or cooperative activity, but to a considerable extent, the people of a given area are still likelier to be in each others' social "pools" than people from other areas are, and marriages seem more common within areas than between them. The areas also serve as a means for categorizing community members according to what are seen as important qualities held in common.

Beginning at Fort Jesus and going northward through the areas (see map 2), the residents of Kavani (including Kibukoni, Mtondoni, and Barani) are spoken of as the richest in Old Town and as having the finest houses. This area includes the largest number of families who view themselves as being of relatively recent Arab descent. They are generally, informants say, seen as educated, refined, and physically unimpressive as concerns fighting for men and hard work for women.

Among non-Swahili Mombasans, a Swahili man is sometimes referred to as Bwana Badi , where "Bwana" (the title meaning "lord" or "sir") is used to mock what is taken to be the Swahili claim to nobility (remember this quality, uungwana, refers to slave-free descent and is a central quality for full membership in the community) and "Badi," the short and familiar form of "Mohammed," a very common name for men in the community. The men of Kibukoni are the ones who are most commonly referred to as "Bwana Badi," including, at least once in my presence, by Swahili from other Old Town areas.

The people of Mkanyageni (including Kuze) are considered the bravest, strongest, friendliest, and most loyal members of the community. They are also considered the poorest, the roughest, the least cultivated, and the most insultingly outspoken and quarrelsome. The people of this area include the newest immigrants (although they have been there for generations) coming from Bajun, the Lamu archipelago, and other Swahili communities along the coast.

The famous fighters of Old Town are all Kuze men, and men from other areas have been heard more than once to say of them (collectively or about individuals), "Walizaliwa na bakora na kisu," i.e., that they are born with a walking stick and a knife. A great fighter from this area who died in the 1940s was said to be able to hit an enemy twelve times with his walking stick and leave a single mark. The most recent murder committed by a Swahili which I am aware of was the stabbing of an outsider by a Kuze man who had been insulted by his victim.

The residents of Mjua wa Kale are characterized as being in between the other two in wealth, refinement, and fighting ability. There are some Omaniderived families in this area, called "the Mjua wa Kale Arabs," who are—or were—associated with the Nine Tribes rather than the Three Tribes as most "Swahili Arabs" (as that term was used in chap. 3) are. The main trait of the residents of this area is thought of as pride expressed as interest in precedence, so, for example, it is said a man from here always refused to follow the pre-

vailing practice of standing when the Prophet's name is mentioned in Maulidi because "nobles don't stand for one another." In a story I heard several times, an Mjua wa Kale man came into a mosque before prayers and was greeted by a neighbor resting on the floor without his hat or gizbau (the traditional men's vest—now rarely seen—worn over the ankle-length kanzu). The newcomer took off his hat and gizbau, lay down on the floor, and, only then, returned the greeting.

Residence Choice and Household Location

My inability to get a complete door-to-door census makes it difficult to say with any certainty whether there is a pattern in newly married couples' actual place of residence and, if there is, what pattern is present. A majority of informants say that in their view, residence for a newly married couple is ideally in or near the groom's father's house. Further, a majority also say that if the choice is between living near the wife's parents or living equally far from both, it is better to live equally far from both.

Despite these fairly generally held ideals, my limited direct observation reveals no clear pattern of residence. Some couples live near the groom's family, some live near the bride's family, and many (seemingly the largest group) live about as far from one as from the other. The situation is complicated by the fact that Old Town is very crowded, and couples often have to live wherever they can find space. It may be that the majority's reported preference for virilocal residence is followed less than it is because people simply cannot find housing where they would like to find it. Whether this is so or not, the best assessment of the facts available to me is that there is no clear pattern of residential choice.

Table 2C shows who lives with whom rather than what household is near what other. It shows that the nuclear family is, generally, the most important domestic unit. Seventy-six percent of our sample involves people living in a household with their nuclear family kin. Almost half the households contain no relatives other than nuclear family kin, that is, no adults other than spouses and unmarried offspring. Since nonrelatives only rarely are part of households,^[11] this means that a large majority of community members live with a spouse and their children or, before they marry, with their parents and siblings.

Still, many households contain kin other than the spouses and their children. About half of the 111 households for which I have census data include at least one adult kin of the person who is said by the members of the household to be mwenye amri ("having authority," i.e., the household head). Widows and widowers, if they do not remarry, usually live in the house of one

of their children (or have one or more of them live in their house). Those without children or without those who can or will take them live where they can, most often in their siblings' or siblings' children's houses.

Married children living with one or both parents together with childless widows and widowers living with collateral relatives account for the roughly one-third of the surveyed households that include either the househead's parent or a parent's sibling.

The average number of children per household was 3.65 for male-headed households and 3.59 for female-headed households. This includes both biological children and children who are lelewa (roughly "adopted"; see below).

"Adoption"

Although reliable census data, including information on number of families with adopted children, are very difficult to obtain, more than half my informants (the men and women I worked closely with over extended periods, perhaps a total of 35) were either themselves adopted ("lelewa" is the passive form, "be adopted," and lea , "adopt," is the active form), have adopted one or more children since becoming adults, or both. Some informants say the practice is "not rare, but not even a quarter of all families do it." Others say as many as a quarter of all families contain one or more children who have been "adopted." The general reticence about private matters is the main factor inhibiting the collection of data about adoption. There is no suggestion that there is anything disfavored about being an adopted child. On the contrary, men and women report that adopted children, if treated differently from biological children, are treated better and loved more.

This is especially so for those children, and this seems to be the larger category, who are adopted as acts of friendship. Many informants report that if two women like one another and one gives birth, the other will request (omba , which can also be used as "beg" is) that she be given the child to raise. If the birth mother accepts the request, and it appears that it often is, the child will be turned over to its mama mlezi (adoptive mother) at forty days, after which it will be brought to its birth mother (mama mzazi ) for breast feeding several times a day for a year or so, when it will be switched to a bottle to supplement solid food.

It is by no means true that all such relationships begin so early in the child's life. Many are "adopted"^[12] much later, and the relationship can be terminated, with the child returning to the biological parents usually because the birth parents are displeased with the way he or she is treated or because the adoptive parents can no longer keep him or her. Such terminations are rare and are taken very seriously by all involved.

The women involved in these relations are, more often than not, kin. Sis-

ters often adopt one another's children, and the same is true for women of the same generation closely related through either their fathers or their mothers. Neighbors also adopt each other's children despite the absence of any kin ties, and cross-generational adoption with a grandmother adopting her own daughter's child, while not common, is not unknown. Men sometimes adopt the children of relatives or even friends, but these adoptions are far more often because of "need" (ya lazima ) than friendship.

Informants all differentiate between two types of adoption: because of need and because of friendship. Men generally instigate adoption when the child of someone to whom they are closely tied is in pressing need of a home because of parental illness or death or other such circumstances. Women are occasionally the main movers in an adoption on such dire grounds, but they are mainly the central figures when adoptions are based on the wish to strengthen relations.

The Lives of Adopted Children and of "Natural" Children

When a child is adopted, he or she continues to see the birth parents on a regular basis and continues to call them baba and mama while using the same terms for the adoptive parents. The visits with the birth parents are just that, visits. The child's home is with the adoptive parents. So far as I could determine, all of the adopted school age boys and girls I know of regularly shinda (spend the whole day as opposed to stopping in for a shorter time) with their biological parents. This type of visit is made by all children, adopted or not, to various relatives. But it is an occasional thing as regards visits to other relatives, while it is often a never missed weekly visit, often on Friday, for those who are adopted.

An adopted child is treated much as a biological child is. In paying the bride-price, for example, the adoptive parents provide the whole amount if they are well-to-do or, at least, a substantial part of it if they are less prosperous. Even as concerns radhi (the parental blessing), many biological parents say that they give theirs if the adoptive parents do and withhold it on the same grounds. I know of adoptive children who provide the funds to support their adoptive parents when the latter are old and unable to provide for themselves.

In general, children leave their parents' home to establish their own household in their early or middle twenties and, as table 2B shows, very few children remain in the parental home beyond their mid-twenties. A few sons, but almost no daughters, leave their parental home before they marry. The sons may leave to take employment in another part of Kenya or abroad, but few sons leave—generally because of quarrels with their fathers—to live with other kin (generally married siblings) or, in unusual cases, in a rented room in the house of a distant kin or unrelated community member.

Marriage for Women, Jobs for Men

The nuclear family, with or without one or two additional relatives, is the household unit for most Swahili. The beginning of these families, marriage, is considered the key event in the individual's entrance into adulthood. There are virtually no women who fail to marry at least once, and only very few men remain single. Such unmarried men are viewed as, at best, impotent and, quite likely, the passive partner in homosexuality (mshoga or hanithi ).^[13]

Women who have never married are considered extremely unfortunate and, until the last few decades, denied the company of married women other than their mothers and sisters lest they hear talk of marital sex and be stimulated to improper behavior. In fact, the young women I have been able to interview are unanimous in saying they want to marry. Without exception, the single girls and young woman I talked to^[14] told me that the most important thing that would happen to her, and the thing she looked forward to most, is getting married. The reason given for this was invariably that in getting married, she would begin her own life, have her own home, and become independent.

Every boy and young man I asked told me that the most important thing that would happen to him in the foreseeable future was to get a job. Just as marriage is said by women to bring them independence, employment is cited by men as the source of their independence. The men say this comes through escape from the father's financial control and with this, as many note, the ability to marry and found a family.

Arranged First Marriages

Informants of both sexes say that their first marriage was or will be in large part determined by the wishes of their families. Repeatedly, this was characterized as a community "custom" (mila or desturi ), or, sometimes, a Muslim custom, that has been followed by many previous generations. The mother and father are identified as the main actors in marriage arrangements, but older siblings, especially brothers, were sometimes said to be extremely influential. I know of two cases in which wedding plans agreed to by parents were affected by objections to the groom by the prospective bride's older brother.

There is no disagreement among informants about the desirability of arranged marriages, and many of them talk as though all marriages, at least first marriages, are, in fact, arranged. This, however, proves to be one of those statements (a "token" in the sense explained in chap. 6) that reflect something other than observable reality. When specifically asked, informants all agree that there are, in fact, children who want to marry someone not chosen by their parents. When the child is very insistent, informants say, it

is better to accept his or her choice than to be shamed by the child running off and marrying without parental approval.

Some informants say that marriage on the basis of the choice of those being married is much more common now than it was even as recently as the 1960s and 1970s, and this may well be so. However, having encountered a few middle-aged and even elderly men and women long married to non-Swahili, I discovered that although a few external marriages are, in fact, arranged, these are mainly with outside men who are either rich or prestigious. Such cases involved Arabs or European men converted to Islam and local women whose parents accepted the proposal of the outsider man.

Such marriages to prestigious outsiders, although not common by any means, seem to have been occurring for several generations at least. Obviously, the families that claim Omani forebears and who claim ties with well-established Swahili mbari (the now mainly forgotten collections of families) through their mothers must have involved extracommunity marriages that took place many generations ago. It seems very likely that these are of the same sort as those now involving rich Arabs and European Muslims.

Occasionally, a marriage is arranged between Muslims of ordinary prestige from outside the community and community women. In all the instances of such marriages involving a woman from a family of waungwana (i.e., a family without known slave forebears) I could find (a total of 11), the woman's family was poor and the man was from another Muslim Mombasa group. In most of these cases (7), the man was ethnically a Hadhrami Arab, but there were cases where the groom was Indian (3) or from another African group (1).

These marriages seem to be faut de mieux . If both families are poor and neither partner has prospects of a marriage that will provide much in the way of wealth or prestige, informants agree that any marriage, so long as it is to a Muslim, is better than no marriage, especially if the union does not involve someone believed to have one or more slave forebears.

All the first marriages involving more prosperous families that I could get information about were arranged and were either within the community or between a woman from the community and a man who was a Persian Gulf Arab, an Islamicized European or American, or from another Swahili group elsewhere on the East African coast or offshore islands.

When the proposed marriage is between kin, the negotiations described below are truncated and more informal than when the proposal and its acceptance involves a nonrelative. If the parents of the couple are close kin (and occasionally they are siblings or the children of siblings), much formality and consultation can be dispensed with since everyone involved knows all the others quite well and since the interests of both parties are much more nearly identical. As the kin relationship between the parents is more and more distant, the marriage decision is more and more like that concerning a nonrelative.

The initiative in arranging a marriage is with the prospective groom's relatives. The decision about which woman should be proposed to (posa [v.] is "propose") is made jointly by the groom's family, with all adult members having a say. The decision is sometimes said to be that of the father and the groom's adult brothers, but all agree that the mother and adult sisters play a key role in providing information about the prospective bride and her family and that their views, especially the mother's, are extremely influential. There are, I have been told, marriages in which the father extended a proposal even though his wife, the groom's mother, disagreed, but such events are said to be rare. Hardly less rare are mothers who convince their husbands, the fathers, to extend proposals the latter strongly resists.

The prospective groom himself often plays an active role in the decisions concerning his prospective marriage even though he usually hardly knows the prospective bride before the wedding. Until the last decade or two, the groom had not normally seen his bride until the wedding night, but now he has often seen her at school and may, in these days of greater mobility and less supervision for young women, know her rather well from meetings at school and in the neighborhood.

Sometimes the young man instigates his family's discussions of his marriage proposal by telling his father and mother that he would like to marry and is particularly interested in the young woman he names who, normally, he has met at school or seen in the neighborhood. The young man's preference is taken very seriously and overridden only if there are compelling reasons for doing so (of the sort indicated below). It is widely understood that young people "these days" (siku hizi , a phrase one hears a good deal when discussing marriage and young people in general) cannot be made to do things they do not want to do, and if the attempt is made, painful consequences (even elopement, a very rare occurrence) may eventuate.

In addition to the prospective groom and his resident nuclear family, other household members (including nonnuclear family kin) are ideally and usually consulted in a serious way, and, in some families, so are senior kin (especially grandparents on both sides and parents' siblings) and parents' siblings who do not live in the household. When kin other than those in the groom's nuclear family are involved, unless one of his parents is missing (through death or divorce), their participation is generally more a matter of form than of substance, although sometimes they will be asked to accompany the father and brothers to the young woman's house when the proposal is actually made. As a friend told me about notifying his son's mother's (my friend's wife) parents of the decision to propose marriage to a particular young woman's family, "We ask them, but really it is only notifying them."

Even though it is almost entirely a matter of form, consulting kin about a marriage proposal is very important to continued peaceful relations between those involved. When people who consider themselves descended from a

common forebear on either side are not notified about a prospective proposal concerning someone they view as a kinsman (or kinswoman after the proposal has been made and is being considered), they may view themselves as having been offended. In one instance I know of, a man—who was generally viewed as rather quarrelsome—was enraged at not being told of a proposal involving someone whom he said was his kinsman until he heard the wedding was about to take place. "They throw me away," he said. "They don't want me because I am poor." He was, in fact, not particularly poor, and one suspects he made that statement to impute the basest motive he could to people whom he considered kin but did not, in his view, treat him accordingly.

Once a choice has been made, one or more men from the groom's family (sometimes including kin from beyond the nuclear family) call on the men of the bride's nuclear family to make the marriage proposal. When the proposal is made by several men from the groom's family, ideally and typically led by his father or, if he is dead or incapacitated, by the young man's adult oldest brother, it is viewed as more prestigious for the bride's family and more difficult for them to reject.

The proposal is made to the prospective bride's father or, again, if he is dead, to the young woman's oldest brother. The prospective bride, as well as the other women of the household, are never present when the proposal is made. The man with authority over the prospective bride has formal authority to accept or reject the proposal, but, in fact, he should and, informants say, always does consult other family members before making a decision. The consultation follows the same lines and concerns the same kin as are involved in the decision by the groom's family to extend the proposal.

Similarly, the young woman's own views are considered by her family much as the young man's are by his. I heard of no marriage initiated by the bride, but her wishes are taken seriously. She cannot get her family to initiate a proposal, but her preference in accepting or rejecting one is never lightly dismissed. At a first marriage, the Swahili believe that, in accord with Koranic law, the bride need not actively assent to the match proposed to her. However, should she explicitly refuse by telling her parents she will not accept the marriage, it will not take place. Although some men and women say that young women only rarely reject the decision of their parents, siblings, and other kin, I know of at least two cases (and suspect there are an unknown number I did not find out about) in which marriages did not take place because the prospective bride did not want to marry the man for whom the proposal was made.^[15]

Selecting the Families of Those Who are to Marry

A central factor in the extending or accepting of a marriage proposal is the view that the two families have of each other's standing as waungwana

(nobles). The suspicion of the existence of a slave forebear, which, as noted previously, bars the object of suspicion from the muungwana status, will almost invariably put the suspected family beyond consideration save by families who are themselves suspect.

The Swahili follow the Shafi canon of Sunni Islam, and that canon, informants tell me, requires that those to be married be kufu . This means that they must be of the same "tribe," the term used for this by English-speaking informants but probably referring in some part to prestige and economic standing. An informant told me that the founder of Shafi said that the child of a religious judge (kadhi) should not marry the child of a blacksmith.

Just all of what is involved in kufu is not entirely clear, since, as seen, Swahili do marry people from other ethnic groups, but community members agree that a proposed marriage can be rejected if the proposed partners are not kufu. At the same time it is also agreed that a rejection on such grounds is the most insulting possible. This suggests that objections based in kufu are at least sometimes used as a way of saying the proposer is not a muungwana, a person of undoubted free birth, or, at least, is of such low station as to be unworthy.

The economic situation of the two families is important to those involved, and although each generally prefers, pari pasu, the other to be equal or somewhat more financially secure, only extreme poverty is a likely source of serious objection to the marriage. The prestige of the occupations has significance in itself, with highly remunerative but low-prestige positions less desirable than less remunerative but more admired ones.

The significance of kufu, equality, is seen in some uses of kin terms for nonkin. Women, usually of different generations, who are fond of one another sometimes address one another as mavya, a reciprocal term for husband's mother and son's wife, even if the actual relationship does not exist. The term, it was explained to me, is a friendly one asserting the equality of the women and the propriety of a marriage that would unite them in the relationship it indicates. Similarly, women fondly call boys of their daughter's age Bamkwe (a contraction of "Bwana Mkwe," where "mkwe" is a reciprocal term for spouse's parents and child's spouse), even if the boys are young children.

Birth and economic status are by no means the only basis for evaluating marriage proposals. The suspicion of serious, unacceptable behavior or character traits on the part of either of the young people will bring the other family to oppose the marriage. If the young woman is thought to have a quarrelsome or difficult character or if her association with the other sex is thought to have been too free, the proposal is unlikely to be made by the groom's family.

Rejecting or Not Extending Proposals

The most common objections leading the woman's family to reject the proposal from an otherwise acceptable family are that the young man drinks

alcohol or smokes bhang (marijuana), has engaged in passive homosexual behavior, is a brawler and troublemaker (mhuni ), or has no economic prospects because of a poor school record or poor work history. Respectful behavior toward others, especially seniors, is considered important for both the bride and the groom, but it is mainly young men who are found deficient in this quality. Informants say that "these days" this proper respect from the young is so rare that even quite disrespectful young men must be accepted if one's daughter is to marry at all.

In one case I am familiar with, the prospective bride's older brother, although absent from Old Town (he had a job in Europe) and not the one seen as having authority over her, nearly succeeded in blocking the acceptance of a proposal because he suspected that the young man (who had an unusual manner) was a passive homosexual. The wedding only took place after the bride's grandmother (her father and mother were both dead), with the help of other relatives including the grandmother's brother, convinced the brother to withdraw his threat to stay away from the wedding and not to contribute to its very substantial cost.

Suspicious that young women have any association with unrelated males, other than the most transitory ones at school, or have excessive interest in sex are taken very seriously. An mkware (a woman with strong and active interest in sex) is considered a splendid mistress but a risky wife. Young women suspected of this quality are by no means favored by families as prospective wives for their sons or brothers. Laziness, however, seems to be the character trait that most often leads members of the groom's family to oppose extending a proposal.

I have been told that a bad reputation for any of the members of the family of the groom can lead to hesitance on the part of the bride's family in accepting a proposal and that, similarly, a groom's family may be reluctant to propose to the family of a prospective bride if there are suspicions about her sisters, brothers, or parents. This is, in part, I am told, because of the difficulties anticipated with the in-laws and, in part, because the prospective bride or groom may not be as desirable as she or he seems but, rather, to resemble the family member of doubtful reputation. I do not know of any cases of such considerations actually preventing a proposal or an acceptance, but such may exist.

The Basis for Successful Marriages

Young men tell me that some brides (I suspect that the number is still quite small) are not actually virgins, but the ideal that they be is very strongly held by the community in general and by both her family and that of the groom.^[16] These same young men say that the groom would mind less if his bride were not a virgin than would his father and the woman's father.

Until recently, the failure of the wedding night to result in evidence of the

bride's virginity (in the form of blood on the cloth that had been beneath her hips on the marital bed) could lead to serious difficulties. From the mid-1970s to the end of the 1980s, no quarreling of this sort came to my attention, as I am sure it would have had it been public. Nevertheless, informants say the groom's family would accuse the bride of promiscuity and the bride's family would insist she was a virgin but that the groom was impotent. In such cases, the marriage would be ended without delay. A few marriages, in fact, do end very quickly, and it may be that some of these are due to difficulties related to understandings about the bride's virginity and/or the groom's potency.

The successful marriage, a number of men and women have separately told me, is based on love (upendo). This despite the fact that marriages are often arranged and that, even in the later 1980s, marriages contracted because the prospective spouses are attracted to one another before the wedding are considered less likely to flourish than those based on family choices. The expressed understanding is that parents and adult family members have sounder and more mature judgment than the young prospective spouses do and that family judgment is likelier to bring together a couple who is truly suited.

When such a couple is brought together, the ideal view holds, they will develop a lasting love for one another. Even initial incompatibility will, informants say, be overcome as the couple gets to know one another. A proverb, sometimes used with regard to people getting used to things quite outside of marriage, holds, Walioana, wataambana (They were married, they will [learn to] speak to one another, i.e., as they stay together, they will learn to get along).

Bride Wealth

The bride wealth (mahari ) has been rising fairly steeply in recent years. In 1980, I recorded a bride wealth payment of KSh 15,000, or around $1,000 at the rate of exchange as it was then, and in 1985, I recorded another one, this time of KSh 32,000, or $2,000 at the exchange rate at that time. The mahari is used by the bride's family to buy furniture and household goods for the newly married couple. The bride's parents ideally keep none of it (I heard of no cases when they did keep even part) and are expected to (in the cases I could follow, actually did) add money of their own to it in order to buy more and better things for the couple's new household. I have been told that the groom can provide furniture rather than money, but I have not recorded an actual case of this happening.

Divorce

The return of the bride wealth when divorce occurs did not occasion serious conflict in any of the cases I have information about. If the furniture

is sold, the groom's parents expect to get the money paid for it, but if it is not, they will accept the chairs, tables, and pans and, even, not necessarily pursue every item bought.

Divorce is fairly common, especially for first marriages. I estimate that somewhat more than a third of first marriages are dissolved by divorce. It is substantially less common for subsequent marriages, and it is not common to find men or women who have been divorced more than twice. Divorce is generally viewed as an extremely unfortunate event that is, however, sometimes a necessity. One commonly hears people say, "Of the things God allows, He hates none more than divorce."

Nuclear Family Life

Nuclear family ties are close ties in all respects. The division of labor is clear and sharp. The mother/wife with the assistance of the family daughters (and, in most even moderately prosperous families, the family servant) cares for all household needs, and the father/husband, generally the only family member with paid employment, is the source of the family exchequer. In some families, there is rent-producing property belonging to either or both spouses adding to family income, and if there are children with paid jobs, they contribute most of their earnings until they marry. Even married children are expected to make some contribution toward the living expenses of their parents and dependent siblings and to help with the costs of educating the latter.

Socially, family members spend a good part of every day together. However, the women and younger children are in the house a good deal more than the men and older boys. The family takes its meals from the common kitchen, though in many families the male and female members do not actually eat together. The wife/mother or a daughter/sister brings food to the husband/father and sons/brothers who eat and leave the table (or, more traditionally, mat) in the living room, at which time the servers eat the remaining food, usually in the kitchen. In some families now, I was told, the whole group eats together, but I have never actually seen this and am told it is not common.

The separation of the sexes, an explicit value, leads to the nuclear family itself being separated for a good part of every day, and this is furthered by the sharp division of labor. Men are never to set foot in a kitchen. Swahili lore has it, a boy who goes into the kitchen after puberty will pay for it by having rain on his wedding day.

Family Activity

The most common joint family activity is talking together (44% of the families surveyed chose this as the main nuclear family activity; see chap. 5).

There is some constraint on fathers talking with daughters, and several informants have told me that until two or three decades ago, it was common for daughters to avoid any face-to-face contact with their fathers. Thus, daughters did not remain in the same room as their fathers and a daughter would only converse with her father by addressing him through the door from outside the room he was in.

Such avoidance is not common now, but there is still constraint so that, for example, when the family television is on, the daughters of the household usually sit together, often with the mother, on one side of the room while sons sit with their father on the other side. The separation of women from men affects not only the daughter but also the mother/wife. A grown, married woman typically and usually has a rich social life with her female kin and neighbors, but her movements are restricted and much of her time is spent at home or visiting in the homes of nearby kin and friends. Women venture out to buy their own clothing and make small purchases from the tiny food and sundries shops (called reshun ) scattered throughout Old Town. The main shopping for groceries, meat, fish, and other household supplies is done by the men of the house.

The sons spend more time away from the family home than anyone else, playing with their male kin and neighbors when they are young and going off to school and work as they mature. The daughters spend a fair amount of time with their same sex kin and neighbors, but they are more restricted in what they do than their brothers are and one rarely sees girls and young women outside their houses.

Both boys and girls attend religious school (chuo, pl. vyuo ) beginning at the age of 5 or so and generally attend for at least a year or two as they learn to read the Koran and to write, usually with limited ability, in Arabic. These schools are timed so that the older children can attend secular school, but many religious schools have long sessions on the days when the secular schools are closed.

Children all shinda (spend the day) with kin in other households, but this is of limited importance for boys who, although they continue to visit, mainly stop spending a whole day at the house of a kinsman when they are old enough to spend their time playing soccer and wandering around the neighborhood with other boys of their age. For girls, the days at the houses of kin are their main opportunity to leave the family home, and they generally continue this more-or-less supervised, daylong kin visiting until marriage.

The Division of Activity by Gender in the Family and Generally

The supervision of girls has become less strict in the 1980s than it was even in the 1970s. Many now go to the same secular schools that their brothers do, and some young women now get jobs in offices and shops after finishing

school and before marriage. A few women continue working after marriage, but almost all stop paid employment when the first child is born. There is still, however, a good deal of concern about girls' activities on the part of their parents and grown siblings. The explicit reason for this is fear that the girl may indulge her sexual appetites before marriage. "Having a daughter," a middle-aged Swahili father told me, "is like having an egg in your hand. You cannot be careless for a minute without it being ruined forever."

Men and boys are free to roam the city as they wish but ideally should avoid bad company and late hours. men and boys attend the mosque, go to work, chat with friends, and some—now a minority but formerly a much larger percentage—regularly attend small-scale men's gatherings (baraza) that generally take place at the same time and in the same location. The Swahili men understand themselves to be very sociable, and whether in a baraza or not, they stop to chat with kin, neighbors, and acquaintances whenever they seen them. There was a very strong value on men "being known" among other men which still exists if, perhaps, somewhat less explicitly than I was told was the case some decades ago.

Men, like their wives, sisters, and daughters, however, are almost always at home at mealtimes. Houses are generally locked up at the completion of the evening prayer (isha, rarely later than 8:30), and the entire family retires early. It is quite rare to see middle-aged Swahili men on the streets after the last prayer, but small groups of Swahili young men can be seen on street corners chatting until 11:00 or, sometimes, a little later.

Other than chatting on the streets, the main activity outside the home is attending cinemas. There are two in Old Town which show mainly American and European films (preferred by many community members who attend films) and another showing exclusively Indian films which some Swahili attend. It is now fairly generally accepted for men and boys to go to cinemas, and some men even go with their wives. Conservatives view this attendance with something between caution and alarm, but that view seems less prevalent in the community at the end of the 1980s than it was at the beginning of the decade and earlier.

During the holy month of Ramadhan, the Old Town area is transformed. The streets, usually deserted in the early evening, teem with men and boys, and the usually darkened, quiet houses are full of music, talk, and life. Following the afternoon prayer (magharibi ) when the sun goes down, everyone eats the first food of the day (futari, traditionally a date) and goes home to prepare for the main meal (daku ), which is eaten sometime after the final prayer (isha) but generally at 10:30 or 11:00. Following the meal, people amuse themselves with games, music, or visiting with neighbors. The streets are full of men and boys chatting or playing checkers or cards (but not for money since gambling is forbidden in Islam), and women's voices ring from the houses. Some families spend this festive time together, but mainly the

men and women celebrate separately. After the brief sleep characteristic of Ramadhan, some families eat a heavy breakfast to sustain them through the day of fasting to come, but others sleep a bit longer and have only tea and whatever snacks are available just before dawn and the reinstitution of the fast.

It will be clear from this brief sketch of leisure activities that men and women, boys and girls have quite different activities: the women's center in the home and the men's outside. Nor is this only a matter of the separation of the sexes. Proper men should go around visiting and be known to the community. Women should stay near their houses and those of their close kin. Women do not veil their faces any longer, although it is only in the 1980s that they have ceased doing so, but they should not be seen in public.

Employment

Save for the more conservative members of the community, it is not understood as wrong for women to have jobs outside the home. But, in fact, few have. In part, this is because of the scarcity of employment combined with a general disinclination to hire women, and, in part, it is due to the restrictions of the job market where few occupations save nursing, teaching, and office work are open to women. Except for baby nurses (aya ), even house servants are generally men, and although Swahili would not consider accepting such jobs, this restriction suggests how limited women's employment opportunities are.

With virtually no exceptions, Swahili men have a job, are looking for one, or are retired from one. This means that while women's lives tend to focus around the home and immediate neighborhood, men's are centered outside the home and, often, include employment that is beyond not only the home and neighborhood but also the community.

The Central Place of the Nuclear Family

For women and their children, the nuclear family is obviously the vital center of their lives. They are together much of the time, cooperate in much of what is most important to each of them, and are united by strong emotional bonds. For men, the situation is less of a piece. Men have few close ties, but the closest of all are with their wives (see chap. 10). Their relations with their children, especially their sons, can be rather distant, and their relations with both sons and daughters are not close in comparison with those between the mother and children. Still, these are the closest relations the men have from

an emotional perspective and the most responsible from a material and social perspective.

We have seen that neighbors and nonnuclear family kin play crucial parts in people's lives, but this does not diminish the role the nuclear family plays. Much of what all group members do is either with other nuclear family members or, as regards men's work, viewed as largely in the interests of those members. Members are in constant contact with one another and are interested in almost everything done or happening to any of the others.

In nuclear families that do not break up through divorce, one would assume that members' activities are guided by similar understandings about the way the family life is and how it ought to be. If the members did not share these beliefs and values initially—as when a couple first marries and as children grow up—there is ample opportunity for mutual socialization, so that the importance of cultural sharing is accompanied by what may be a unique opportunity to achieve that sharing.

Thus, the cultural elements concerning this group, it would seem, are at least as likely to be shared among its members as most others. In the following chapter, the extent of this sharing will be examined and it will be shown that it is, in fact, substantially less than complete.