1
Ethnographic and Theoretical Introduction

This is a book about culture and how it actually works in guiding the behavior of those who, in the broad sense, share it. It is also a book about the Swahili of Mombasa and how their culture operates to guide their social lives and to provide them with a means for dealing with the problems and opportunities they encounter. The aim is to contribute to our understanding of the processes whereby culture works for all humanity and, specifically, to examine its constituent processes as they are seen among the Mombasa Swahili.

Ethnographically, field work focused particularly on interpersonal relationships, especially marriage and family life, generational relations, the ties among neighbors, and community structure. The social aspects of shame and the beliefs and practices concerned with health and illness are given special attention from the perspective of their cultural foundations. Taken together, they provide a reasonably broad and inclusive ethnography, despite the fact that they were chosen as much because of their theoretical interest and the availability of information in a community where information is remarkably difficult to obtain.

In examining the various aspects of Swahili life, the main goal is to identify their cultural bases and understand the processes whereby culture guides the behavior of community members to produce the social phenomena observed. The central thesis is that culture's functioning is best understood from a perspective that puts particular emphasis on the part played by a pervasively important combination of cultural elements, "status." This cultural complex will be shown to be central to the fact that culture's effects are not limited to those who share the cultural elements in question. It is through statuses,

in fact, that the crucial processes of distribution, organization, and differential promotion of conformity will be shown mainly to operate.

The Ethnographic Focus

As will be seen in chapter 2, the Mombasa Swahili are part of an ancient urban community that has been in its present location on the coast of what is now Kenya for centuries. The members of this group view themselves as the heirs to cultural traditions that remain vital guides to behavior despite changes in their community and in the city their forebears founded nearly a millennium ago.

Part of the group's tradition is seen in the two-section organization of the community. As chapter 3 shows, in recent decades, this community has been strained by a weakening in the division between the sections through individuals claiming statuses that would place them outside the community and unite them with others from whom they were previously separated. This strain has been intensified by what are seen as claims for community membership from occupants of statuses that were not formerly understood as members. These strains have diminished the community's integration and stopped most joint activity. They have not, however, undermined the community's effectiveness as, in many senses, the arena for its members' lives. It still provides its members with the cultural foundation for living and the social framework within which they are born, work, marry, raise children, and die.

Chapter 4 examines the Swahili nuclear family and shows that in this largely endogamous community, it is far the most significant grouping in its members' lives. Kinship beyond the nuclear family is quite important, and ties with neighbors are lasting, but it is in relationships with parents, spouse, children, and siblings that most community members spend most of their time and much of what is vital to each person takes place.

Despite this importance, chapter 5 shows that even in this effectively functioning community, the sharing of cultural elements concerned with some of the fundamental issues in nuclear family relationships and group concerns is strictly limited. Members of long-established and stable nuclear families were interviewed concerning nuclear family issues (e.g., "Who makes decisions in your house?") and values ("Should children love their fathers more, their mothers more, or both the same?"). Informants' responses were compared with their fellow family members, with members of other families who occupied the same family statuses, and with all other informants without regard to family membership or status.

This study showed that even in the groups with the highest level of sharing, that is, among members of the same nuclear family, more than a quarter of the items were not shared and that within the community as a whole, almost

a half were not shared. It was also found that individuals belonging to the same status, for example, "daughter," shared the cultural elements concerned with that status less with other occupants of that same status than they did with those who shared with them the status "member of my family."

Since the nuclear family among the Mombasa Swahili is a co-resident group whose members spend a great deal of every day together, since marriage in this group is mainly endogamous to the community, and since no questions were asked about matters beyond the scope of the nuclear family's life, it seems a reasonable working hypothesis that cultural sharing in other social settings (i.e., outside the nuclear family) concerning other issues is unlikely to be much greater save, perhaps, in the area of technical knowledge shared among those in the status devoted to its employment.

There is some basis for believing that, in fact, there is less sharing in other areas of life (as work by Fernandez [1965, 1982] on ritual, Keesing [1987a ] on eschatology, and Holland [1987a ] on academic matters among students suggest) than within the nuclear family. However, even if sharing is as great as within the family, the probability that it is less than complete in all relationships and concerning all issues is, unless specifically shown otherwise, taken as a basic element in discussing culture's functioning in the highly integrated Swahili community.^[1]

The fact is that the social lives of the members of the community are, aside from limited relations based on schooling and occupation, almost entirely within the community. Further, membership in the community is an unquestionably important part of the identity of every one of the scores of members I have talked with over the years, and the ethnocentrism to be expected in a functioning community is decidedly present. Taking these facts together with the nearly endogamous patterns of friendship and marriage shows that the group's culture remains vital and effective.

Incomplete Sharing and Cultural "Explanations"

The fact that there are demonstrated limits on the sharing of culture in this community's most closely associated group, the nuclear family, provides a well-marked opportunity to study culture's functioning with the invocation of "shared beliefs and values" clearly an insufficient explanation of what is observed. The basis for such an invocation has been removed by a series of studies showing culture's elements to be only partially shared (see Roberts 1951; Wallace 1970; Schwartz 1978; Willis 1972; Pelto and Pelto 1975; Swartz 1982; D'Andrade n.d.; Holland 1987a ; and others). How culture works despite its incomplete sharing, including incomplete sharing within statuses, is only beginning to be investigated (e.g., Gearing 1976a and 1976b and Holland 1987b , for a mainly cognitive approach).

Even if culture were fully shared by everyone or, at least, fully shared within particular status groupings, the nature of the relations among its parts and the sources of its effectiveness would still call for close study. The dynamics of culture, the processes whereby it guides the behavior of individuals and serves as a foundation for social relations, have never received much attention beyond broad characterizations such as their being controlled by evolution, diffusion, or environmental adaptation.

"Molecular" Processes and the Enduring Myth of Complete Sharing

The grand processes of cultural development have been, and continue to be, of great interest, the perspectives being the sweeping, universal ones associated with such writers as Marx, Toynbee, and Weber. But the everyday processes whereby culture actually accomplishes what anthropologists say it does, that is, provide the basis for the distinctively human mode of adaptation, have received far less attention. Cognitive anthropologists are currently in the forefront in this molecular approach where those interested in culture and personality once led and still make important contributions. These studies, however, mainly limit their focus to the psychological aspects of cultural dynamics. The study of the social aspects is still surprisingly neglected.

A main theoretical basis for that neglect is the enduring view that culture's part in individual adaptation and the regulation of social relations can be accounted for by reference to "shared beliefs and values." Although explicit support among students of society for the view that any culture is universally shared has vanished, much of the general understanding of how culture works is based on the implicit position that if a belief or value is truly part of culture, that is, if it is shared by group members, the invocation of that belief or value is sufficient to explain social phenomena.

Never mind that the invoker has usually granted that all beliefs and values are not, in fact, shared by all group members. Questions of cultural distribution, of who shares what with whom and how cultural elements find their way to the situations where they are used, are only rarely raised (Schwartz [1972, 1978] has been a leader in broaching these issues), and the questions concerning the relations among culture's elements are far more often left aside than considered. The issue of how cultural elements affect those who do not share them, clearly of central importance to culture's functioning if its elements are not uniformly shared, is almost never examined. The same is true concerning how cultural conformity is encouraged given the diversity that is part of incomplete sharing.

The tacit acceptance of completely shared culture, despite an avowed rejection of the view, makes it possible to ignore these and related issues, al-

though the cost from the perspective of understanding the functioning of the human adaptation is substantial. Here the interest is in the dynamics of Mombasa Swahili culture at what might be called "ground level." Explicit attention is directed to cultural distribution, organization, and differential conformity from a perspective that totally forswears the invocation, tacit or otherwise, of any but demonstrable sharing. This perspective is intended to contribute to raising new and, sometimes, different questions aimed at advancing understanding of fundamental cultural processes.

Culture and "Culture"

Culture, as it is understood here, is not the only source of influence on human behavior, but it affects everything people do and is the indispensable base for social relations. There may be some merit in the vertiginous, Weberian metaphor wherein humans are suspended in a web of meaning whose substance is culture, but the trope distorts the realities of human life.

Unless "meaning" is understood so broadly as to be almost useless in analysis, Weber's web is spun of only one of the two broad contributions culture makes to human life. The other is to provide instructions for doing things such as making money, friends, and love; what I call "procedural understandings" (Swartz and Jordan 1980:49) and in some ways similar to what Goodenough calls "recipes" (1971:30–31).

"Culture" is a concept referring to the propositions or understandings that exist in the minds of individuals and that are similar to the comparable propositions in the minds of their fellow group members. That is, the understandings that make up culture are shared, so that culture's locus is both social and psychological. The basic units of culture are in people's minds, but the fact of their being shared, manifested and influenced in interaction, is a social fact. The component shared understandings have both direct and, as we will see below, vitally important indirect influence on the members of the group, some of whose members share them.

The direct influence comes from the guidance the understandings provide both for behavior and for the evaluation and assessment of whatever the actor views as relevant to that behavior. This direct influence is psychological in origin, and its primary operation is in the cognitive processes of the culture sharers. But culture's effects are not limited to those processes, crucial though they are, that occur in the human mind.

The manifest, public activity that is guided by culture is not itself culture, but it is, in considerable part, produced by it in the sense that the behavior of those involved is guided by shared understandings. This behavior, when assessed and evaluated according to understandings shared among those who express it and those who are concerned with it, is itself an influence on the

subsequent behavior of those involved and, often, on the behavior of those who become aware of what was done. The behavior produced under cultural guidance thus serves as a vehicle for further, but indirect, cultural influence (see Goodenough 1971:18–20).

Some of the phenomena addressed in my approach, including the simultaneous occupancy of a number of "statuses" (in the sense explained below) and the importance of different sorts of situations to how statuses function as guides to behavior, are interestingly and differently developed in Goodenough's pioneering work (1965:12) as well as in Keesing's Chomskian-ethnoscientific "building blocks model" of social participation and cultural competence (1970, esp. 432–436).^[2] I did not have the advantage of reading Keesing until the field research on which this book was based was completed, but I am struck by the thoughtfulness of his formulations and the fact that many of his suggestions work well with the data presented here.

"Status": Culture's Action Arm

The heart of the approach here is what I am calling "status," which is nearer Keesing's use of "role" than any other single concept in his or Goodenough's work (see Keesing 1970:427). The basis for my usage, as for Goodenough's and Keesing's, is Linton's unelaborated notions of "status" and "role" (1936:113–115), but my use of the concepts is much modified and has benefited from the work of the others.^[3]

Clearly, an approach to culture's effect on behavior that seeks to examine the complete range of relevant phenomena must include provision for considering the ways behavior is affected by the products of culture as well as by the direct effects of culture itself. As will become clear in the course of this book, especially in chapters 9 and 10, although statuses are complexes of cultural elements and nothing but that, their operation in guiding behavior has vital effects on behavior that are independent of culture's direct influence. In other words, statuses guide behavior, and that behavior, a product of culture but not itself culture, has its own effects.

Statuses are uniquely important to cultural dynamics. They are what might be called "the action arms of culture." Not all of the understandings shared among the members of a group are parts of statuses, but most of them are. Only speech rivals status in the breadth of its influence on life; not even technology surpasses these two potent culturally based sources of influence. Sociolinguists have developed a thriving inquiry into the effects of speech on social life and culture itself, but status has not been so thoroughly studied.

To show how status comes to have so central a part in influencing behavior, it is useful first to consider what exactly "culture" refers to.

Culture and Its Distribution

"Culture" as used here refers to all the understandings that are socially learned and transmitted and that are shared by two or more actors who consider themselves to belong to some common grouping. This definition is closely related to those used by D'Andrade (1987:195), Goodenough (1971:22), Keesing (1970:440), Frake (1962:85), and Spiro (1984:323), to name only some. By confining "culture" to understandings, the concept not only focuses on mental processes but also only on those that are cognitive. It is not that emotions and perceptions have nothing to do with culture, far from it, but rather that they are not definitionally part of it.

It is worth noting that in this definition, cultural elements need not be shared by all the members of a group. The understanding that they belong to the group will always, by definition, be shared by virtually everyone, and some other understandings may, as a matter of fact, be widely shared, but such universal sharing is not part of "culture" as defined here. So long as two people who consider themselves to be members of the same group share an understanding, it is included as part of the culture of that group. Generally, it appears that each group member shares some understandings with some group members and others with others. Complex and overlapping networks of incomplete sharing of particular cultural elements are what is characteristic of human groups.

In fact, the differential distribution of cultural elements among individuals and categories within a group is itself one of the key influences on behavior within the group as a whole. This fact needs to be emphasized by explicitly including as cultural those understandings, and they are a very large part of the total, that are shared by only a limited part of the whole group's membership. The contents of culture, the famous "beliefs and values" that group members actually share, are what students of society and behavior have most attended to. But a focus on what everyone shares slights the vital importance of the distribution of culture, which is itself, independent of the contents involved, a significant part of culture's influence.

Status: The Action Arm of Culture

To appreciate this, it is essential to understand that statuses are the instrument through which cultural distribution takes place. It is through statuses that it is established which cultural elements are associated with which individuals according to the categories they are understood to occupy in different circumstances. It is also through them that culture comes to bear on the problems and opportunities of life through particular understandings being

associated with particular people according to the statuses they assume and are assigned in various situations. Statuses, then, are the core of culture's part in social life. Although statuses are at the heart of culture's ability to serve as the basis for group life and individual adaptation, they are themselves culturally constituted. All of the components of statuses are cultural elements that can be divided, for purposes of analysis, into three different kinds of understandings distinguishable according to their different functions.

The logically prior of these three kinds of elements are the "identifiers" that serve as the basis for recognizing individuals as members of different categories. Another component, "expectations," indicates how category members are expected (by themselves and others) to behave and react both in general and in particular kinds of situations and contexts. The third type of status component, "salience understandings," concerns the appropriateness and relative weight accorded different category memberships in various situations. Since everyone classifies himself and is classified by others in what is invariably a very large number of statuses, salience understandings serve an essential function in indicating which one or ones are called for and acceptable in what circumstances. Since people often serve in more than one category at a time, salience understandings serve to indicate not only which understandings apply but also what priority is given particular statuses when more than one is called for in a single situation.

Identifiers

I agree with Goodenough (1965:2) and Keesing (1970:424) that the understandings that define status category membership may be different from those concerning the behavior and responses expected from category members. This can be seen, for example, with respect to the question of who is properly categorized as a Swahili. The differences between a "Swahili" and an "Arab" living in the Swahili neighborhood have long concerned the members of this community.

Controversy and shifting social alignments focused around this status assignment have been going on since the turn of the century and before. In fact, the question of who is and who is not a Swahili is, as shown in chapter 3, a major source of the change in social life that has taken place in Old Town in the second half of the twentieth century. Despite its importance, within the group itself, the status "Swahili" or "community member" is an unmarked category without a generally used name. One never hears the words Mswahili or Waswahili , the singular and plural of the noun that would be used to refer to group members if such a term were used.

There are, as we will see in the historical and ethnographic sketch given in chapter 2, specific traits, more or less directly observable, that members of the Mombasa Swahili group all share including, notably, being Sunni

Muslims of the Shafi canon. But the fact that this and other traits are widely or, even, universally shared by community members does not necessarily make them either sufficient or necessary as the basis for the understandings that lead someone to be categorized as a Swahili. Thus, religious affiliation and belief is not enough to serve as the basis for classification by itself, since there are followers of the Shafi canon (including many whose forebears fairly recently migrated to the city from the Indian subcontinent) who are universally understood by community members as belonging to other groups.

Further, in addition to being not sufficient for membership in the unmarked "Swahili" category, there is a case that suggests that religious affiliation by itself may also not be a necessary characteristic. A Swahili poet from Lamu, Mohamed Kijumwa, went so far as to convert to Christianity—among many other behaviors viewed as bizarre and outrageous—but even this did not affect the fact that he is unquestionably viewed and evaluated as a Swahili.

In the case of Sh. Mohamed, the absence of an identifier, being a Muslim as all other Swahili are, did not block him from being categorized as a Swahili and being subject to the expectations associated with that status, so that the two types of understandings are seen to be partly separate. Sometimes, however, the understandings that provide identification are the same ones as those that serve as expectations. Thus, all those who give the prayer call (adhan ) from mosques are muadhini , and no one is who does not. Nevertheless, even in these cases, more is gained in analytic ability by saying that the identifiers and the expectations are part of the same construct than by keeping them entirely separate, as Goodenough's "rights and duties" and "social identities" approach does (1965:3–18).

There are immigrants, and the children and grandchildren of immigrants, from Oman or elsewhere in the Persian Gulf who live in Old Town. Most of those who are the children of immigrants and all of the grandchildren speak Swahili, and, of those, some have converted to Shafi from the Ibathi canon followed in parts of the Gulf region. If these men (I have no data on women immigrants of this sort, if, as is doubtful, there are any) associate with undoubted members of the Swahili category most of the time and smoke and shave their beards (Ibathi do neither), they are mainly evaluated and treated according to expectations that are part of the Swahili status.

But there is a difference in how they are treated and expected to act even if it is a subtle one. Now and again, one hears one of these immigrants referred to as Muarabu (Arab, Warabu pl.), and it seems likely that they are assigned to the Arab status category as well as the Swahili category and that the expectations focused on them include those associated with both. When they behave in ways that are in accord with understandings concerning proper behavior by Swahilis, they are mainly evaluated and reacted to as Swahili by many community members despite a personal history that is not entirely Swahili (because it involves fairly recent immigration). When they do not be-

have in accord with understandings that apply to Swahili, many community members treat them according to Arab status expectations.

Salience Understandings

There is nothing unusual about two or more of an individual's statuses being involved in interaction at once. That is, salience understandings include those that allow (or, even, require) that acting and being evaluated according to one status involves simultaneously acting and being evaluated according to another.^[4]

The Swahili, to take one example of such salience understandings, differentiate among themselves on the basis of neighborhood. It is generally understood that all forms of manual labor are inappropriate for members of their group, with the partial exception of commercial fishing, which is practiced mainly by the men from the Kuze neighborhood of Old Town. These men are unquestionably classified as Swahili, but, unlike other Swahili, they are understood to be more blunt, uninterested in elaborate etiquette, and generally more direct and physical in what they can be expected to do.

This Kuze status is quite as real as the Swahili status, and both have expectations concerning their members. Since the Kuze men meet the identifying understandings for the Swahili status, they are classified that way, and since they also meet the identifiers for Kuze, they are classified that way as well. In some contexts, the Swahili status is dominant; in others, the Kuze status is. This can be seen in their own behavior and that of those interacting with them through inferring the expectations involved.

The occupancy of several statuses by each of the participants in a single event is quite common, and these multiple occupancies are often simultaneous, with the different statuses contributing differently to the expectations that guide the behavior of the statuses' occupants and of their social partners. Gender-based statuses, age-based statuses, and ethnic-based statuses often occur only or mainly in conjunction with other statuses, and, among the Swahili and others, almost all other statuses are occupied jointly with all three of these. But it is not only broad and widely inclusive statuses that are jointly operative in interaction.

The Swahili fisherman is a businessman when he sells his fish. These two statuses with their quite specific and limited expectations are occupied simultaneously, perhaps, with the broader statuses of Muslim, Swahili, Kuze resident, male, and person of whatever age. That this is so is discovered by inferring the expectations of those with whom the fisherman interacts and his expectations of them and comparing these to the expectations found in the various individual statuses at issue.

It would be possible, of course, to speak of a single status, businessman-fisherman or Kuze-Swahili, rather than a combination of the separate statuses.

However, so long as the understandings involved sometimes occur independently in association with only one of the sets of identifiers (e.g., so long as there are understandings about fishermen independent of businessmen), the analysis is better served by treating the statuses as occurring jointly.

Specific and General Expectations

Statuses differ in their importance both to the individual and to the group, and this difference is often associated with a difference in the kind of expectation they have. The differences can be seen in everyday life, of course, but they show themselves clearly in the attacks group members make on one another when angry. A particular sort of aggressive speech, "badtalk" (i.e., speech generally considered rude, coarse, and obscene), aims at questioning the targets' worth as assessed according to the expectations in the most fundamental statuses involved in group life (Swartz 1990a , 1990b ).

The most pointed attacks concern statuses such as community member, the true child of particular parents, or proper male and the relationships involving them. The attacks are mainly in terms of general, rather than specific, expectations. Statuses involving more specific expectations in relationships, such as those in being a bus conductor in relations with passengers, are less broadly important to social participation and usually psychologically less vital. These are not attacked by badtalk as the statuses involved in more crucial relationships are. The most common badtalk used, mainly by women and young men, is, "Your mother's cunt!" implying an improper parent and an improper relationship with inappropriate expectations of so broad a scope they need not be mentioned. No one is assailed with "Your bus conductor's cunt!" because the conductor status, as such, involves only superficial relationships with quite specific expectations.

Specific expectations involve narrow ranges of behavior that require little interpretation. The conductor either gets the fare or the passenger is put off the bus. General expectations, such as children loving their parents or friends helping one another, are quite different and call for elaborate interpretation of behavior in assessing whether they have been met. The reference to the mother's sexual organ is an assertion that the target of the badtalk fails to meet the expectation that he or she be the proper child of particular parents, which is made doubtful by having a mother so sexually active that her most notable characteristic is her vulva.

General expectations are broad and rather vague, so that being accused of failure to meet them is always possible. One cannot successfully accuse a tall person of being short, but everyone can be accused of having a mother whose chastity is less than it ought to be. But the opposite is also true. General expectations can be taken as being met by a broad range of behavior, not just a few specific acts. Most general expectations concern behavior by the status

occupant himself, rather than by his parent or other connected person, but that behavior is always in need of interpretation far beyond what is called for by specific expectations.

General Expectations, Cultural Sharing, and the Effectiveness of Statuses

Specific expectations are part of the substance of everyday life and must rather often be met if social life is to continue. Knowing that a butcher will accept money in exchange for meat, that those next to you at the mosque will follow the imam's example much as you do, and that a greeting will be returned with either no response or a benign one are all important to those who participate in Swahili social life.

Such specific expectations, however, are only a fraction of what is called for and supplied by social relationships in every enduring society. General expectations must also be met in their own ambiguous way. In Swahili society, for example, children in relationships with mothers expect to be helped and cared for, husbands in relationships with wives expect support, and wives in relationships with husbands expect to be shown love.

What behavior demonstrates help and caring, being supportive, and showing love is far more ambiguous than what is involved in the specific expectations concerning buying, greeting, and praying, but the general expectations are by no means less important because of that. The fact that meeting general expectations depends at least as much on interpretations as on behavior itself gives social life a flexibility and tolerance for differences it would otherwise not have. This flexibility and tolerance serves not only to reduce the need for detailed cultural sharing but also to make relationships effective in new circumstances and domains. The limited cultural sharing found among members of functioning nuclear families described in chapter 5 is partly to be understood as related to the latitude in relationships as a consequence of the importance of general expectations in them.

"Tokens" and "Guides"

The ambiguity of general expectations is by no means the only element in culture's ability to function as a guide for social relationships and individual behavior despite many of its elements being shared by only a limited number of those who are affected by them. Another is rooted in the fact that for some purposes, it is as important for those in a relationship to believe they share some understandings as it is to actually share them. Predictability is indispensable to social relations in that if people do not feel relatively confident that

they know the limits within which others will respond to them, they are unlikely to participate in the relations.

Since members of even groups with the most intense and frequent relationships do not share many of the understandings about those relationships, the basis for predictability cannot be assumed to be the "shared beliefs and values" often invoked to explain social phenomena. In fact, as chapter 6 shows, the appearance of sharing even in the presence of manifest differences can serve to facilitate the predictability essential to social life.

This "appearance" need not be a false one. For some purposes, it is useful to distinguish the functions of cultural elements according to whether they actually guide the behavior they most directly concern or are presented mainly as reassuring symbols to others whom the presenter believes use them as guides. A study of intergenerational relations shows (see chap. 6) that in formal interviews, younger Swahili say that the problems between parents and children and seniors and juniors are mainly due to their, the younger people's, lack of discipline and respect. In these same interviews, older Swahili say that their failure to be firm is the major cause of the intergenerational difficulties.

In more informal settings, however, quite different views appear: the younger people blame the older, who, in turn, blame the younger. In most contexts, it is these latter views, rather than the self-blaming ones, that actually guide behavior. In many social situations, young people tell each other about their parents' (and, in general, elders') conservatism and failure to understand the modern world, and they adopt various behavioral strategies to avoid the consequences of these understood attributes of their seniors. In a complementary way, the older people sometimes comfort each other about what they characterize as the abandoned and irresponsible behavior of their children and the hopeless nature of the younger generation. In fact, much of their behavior concerning young people can be seen to be affected by these last understandings even though that is not what would be expected on the basis of what they told the interviewers.

The understandings suggested by statements in the formal interviews of informants from both age groups are real enough; they do exist in the minds of those who state them. They do not, however, provide much actual guidance for behavior in the relationships with those they concern. The interview statements can be called "tokens," as distinguished from "guides." These tokens are intended, and often serve, to reinforce the belief that the individuals using, often exchanging, them are sufficiently similar to be able to predict one another's behavior.

"Cultural Models" and General Expectations

Another contributor to the effectiveness of nuclear families is connected with the effectiveness of the "cultural models" discussed in chapter 7. These

are symbolic representations of desired or despised behavior that are frequently presented or exchanged in emotional and value-laden ways. They gain force from the fact that they are not general or abstract statements of abstract values and ideals. The terms and characterizations that constitute the models are usually quite unambiguous in either praising or condemning identified actions or sorts of behavior carried out by particular individuals in specific social relationships. The models, therefore, provide quite clear indications of what is positively evaluated behavior for particular kinds of individuals in particular settings and what is not.

The amount of sharing needed for the understandings that constitute statuses to be effective as guides for behavior, then, is reduced by lessening the ambiguity concerning what the expectations associated with the statuses actually call for. Further, the models not only make clear what behaviors are well received but do so in a way that promotes those behaviors by presenting them in an emotionally charged and value-laden way.

Goodenough and Keesing present pictures of society in which what I call statuses are occupied by individuals who are taken as sharing with their status mates the understandings that are central to those statuses. For these two theorists, many of the most crucial issues deal with how these shared understandings are related to and interact with one another. Their main concern is with developing a "grammar" of culture using statuses as a key means for discovering the rules. This, however, is contrary to the data presented in chapter 5 as well as the data in Keesing's own 1978 study. If, as these data suggest, culture is quite imperfectly shared even within statuses, their linguistics-based model is not as compelling as it otherwise would be. In language, people employ a vocabulary whose more obvious references, at least, are shared and whose use is according to universally shared rules.

If statuses are taken as central to society's operation and if those statuses actually depend on relatively complete sharing of their elements among those categorized in them, it seems reasonable to infer that in a society functioning well enough to continue, the more central statuses are to relations and group activity, the more the understandings central to those statuses' functioning will be shared. The Swahili data, however, indicate that this is by no means clearly so.

Specifically, the data in chapter 4 show the nuclear family in this community to be functioning at least adequately, but, despite this, cultural sharing within family statuses is distinctly limited. This limited sharing is most clearly characteristic of the understandings about what specific sorts of things status occupants can and should do and how they can and should do them.

These data do not contradict the importance of cultural sharing within statuses. The necessity of having some actual and dependable sharing remains, but that necessity brings us back again to the importance of general, as opposed to specific, understandings. Statuses are quite as important to cul-

ture's functioning as Goodenough and Keesing have maintained, and they do function by distributing cultural elements among group members and, mainly through their salience understandings, among situations. A key to their ability to do this is not complete sharing within statuses, however, but the distinctive quality of general expectations concerning different statuses augmented by the efficacy of readily displayed cultural models that teach and encourage behavioral conformity. What is shared in the most limited way are specific understandings, while those with less easily identified behavioral manifestations are more broadly held and more central to the functioning of the sorts of statuses central to nuclear families and other intimate groups.

Still another source of the efficacy of statuses is found in the apparently thorough sharing (discussed in chapter 4) of status identifiers. The agreement about the assignment of people to statuses seems higher than many other sorts of agreement. There are fairly unambiguous, explicit understandings about status assignment of the "stethoscope wearer = doctor" sort. But such explicit understandings are not necessary in many reciprocal assignments, since the process is one that works itself out in interaction. In actually relating to one another, participants in interaction each adjust to the partner until agreement is reached about the category to which the other is assigned. This is a failure-proof process in that when agreement is not reached, relationships are ended before they really begin.

Identifiers, General Expectations, and More on Models

In communities like the Swahili where a limited number of people spend their lives in the same small area, status assignment is even less fraught with difficulties. When the various status memberships of those around one are well known, the meeting of expectations is simplified. Since the statuses important to much of life in such a community have mainly general, rather than specific, expectations, only limited sharing (in a quantitative sense) is needed for the relationship to proceed fruitfully, that is, the participants meet one another's expectations sufficiently often for them to continue.

There are, in fact, culturally constituted means for encouraging the sort of limited conformity that is called for to meet general expectations. To be competent as judged by peers, a surgeon has to conform quite closely to the technical understandings concerning the work associated with his status, but for a man or a woman to have haya (modesty and concern for the rights of others) calls for no such close agreement between actions and shared understandings. By praising haya, conformity to a variety of distant expectations that are important to the evaluation of individuals in a considerable variety of statuses is encouraged. I will show (chap. 7) that there are a considerable variety of terms in the Swahili language that encourage conformity with the

distant expectations of a range of statuses by providing emotionally potent models of the virtues of conformity and the costs of nonconformity.

It is not that status members in vital statuses such as "mother," "respected man," or "proper child" are led to conform to direct expectations in what they do and how they react but rather that they are encouraged in such general and diffuse conformity as these distant statuses call for.^[5] An examination of the Mombasa Swahili terms most applied to admired and, more often, disapproved behavior shows that not only are these terms almost always general in their reference but they are also almost always concerned with social relationships.

Thus, for example, everyone is concerned with fakhri (which can be glossed as "honor"), and it is a main dimension of evaluation and assessment both of how a person acts and of how that person is treated. How people in different statuses, particularly men as contrasted with women, get fakhri and behave to show fakhri, however, depends on the different statuses they occupy. The ways group members use such value-laden terms exerts a pressure for behavior in conformity with expectations that are status specific and, at the same time, of the distant sort that is met by a considerable range of behavior, lessening the need for a full and detailed sharing of understandings between actors and evaluators.

"Shame" and Its Agents

In chapter 8, shame (aibu in Swahili) is seen to operate in some ways similarly to the terms concerning social relationships. On the one hand, there are a small number of specific understandings that are used as the basis for evaluating all group members as such, regardless of the other statuses they occupy. These universal values are closely associated with the community's most respected and prestigious men whom I call "the arbiters." The concern with being positively assessed with respect to these understandings leads to considerable homogeneity of behavior in certain public domains despite the undoubted differences that exist among those involved. The specific understandings that lie behind these instances of conformity (e.g., that public nakedness is possible only for psychotics and is deeply shameful even for them) are universally shared and almost never violated.

An important thing about this limited but public and ongoing conformity is that it implies a broader and more general agreement concerning understandings than may, in fact, exist. By following a few, limited understandings whose behavioral manifestations are obvious, group members reassure one another that they are quite similar in the things that matter most.

This reassurance is important because, in fact, group members differ both in the understandings that guide their own behavior and in those that serve as the basis for evaluating that of others. By following a few universally

shared understandings, the implication is transmitted that similar agreement exists as concerns the foundations for all behavior. In fact, a far more substantial number of understandings than the few that are universally shared serves not only as a guide to one's own behavior but as the basis for judging most of what others do. Which of this larger number of understandings is used to guide behavior and assessment of others' behavior depends both on the statuses of the evaluators and the statuses of the evaluated. It is fairly obvious that how an individual acts toward another depends on both of their statuses. The same is true of how individuals evaluate one another.

Everyone in the society serves as what I call a "sanctioner," and the judgments of these sanctioners depend on the relationship between them and the individuals who are the targets of their evaluations. This leads to the same behavior being quite differently evaluated depending on who performs it and who is evaluating it. The evaluated suffer shame if they are negatively judged by either arbiters or sanctioners, so that shame serves, as the values attached to and expressed in relationship terms do, to encourage conformity both to generally held and universally applicable understandings and to those that apply quite specifically to particular statuses.

It is important to note that the understandings underlying evaluations by sanctioners are often general rather than specific. The broad and nonspecific nature of the understandings used in these judgments limits the amount of sharing required and, at the same time, encourages flexibility. Further, since these general understandings are part of statuses, they apply differentially to the same individuals as they move from one situation or relationship to another with the changes in statuses such moves often entail. This protean applicability of general understandings promotes both the maintenance of status differentiation and the distribution of culture among actors and across situations while requiring a minimum of sharing of specifics.

"Role" as a Part of Status

The usefulness of the status concept is increased when a distinction is introduced between "status" and "role."^[6] A role is a subunit of the set of understandings that constitute a status including only the understandings concerning relations with others according to the status those others occupy. Thus, if one takes the "arbiter" status, an examination of the behavior of its occupants acting in that status (as indicated by their meeting its identifying understandings) will show that different understandings are involved in guiding the behavior of arbiters in "arbiter-arbiter" interactions as compared to those involved in "arbiter-junior person"^[7] interactions. The "arbiter" status, then, is seen to be involved in two distinct roles and may be involved in others provided only that membership in the arbiter, rather than some other, status is what is salient for at least one participant in the relationship.

Turning to arbiters' function in promoting conformity and group operations, it is important to understand that the individuals categorized together as what I am calling "arbiters" rarely, in fact, can be seen to make judgments. As a matter of fairly explicit policy, those categorized in this status almost never say or do anything that might indicate what their assessment of acts or individuals may be. They serve, in fact, as a sort of culturally constituted Rorschach; they are the embodiment of the famous "they" who appear in the "what will they think" heard in many societies.

Chapter 8 makes clear that the arbiters' imagined judgments are not as frequently of concern as are the judgments of what I call the "sanctioners," but the arbiters play a central part in promoting conformity nevertheless. The sanctioners' judgments are concerned with detailed and specific aspects of what one does in quotidian statuses such as spouse, neighbor, or fellow employee, while the arbiters are mainly taken as concerned with one's standing as a group member, man, women, or human being.

General Expectations and the Effects of Unshared Culture

To this point, attention has been directed to the nature of Swahili culture, the extent of its sharing, and some fundamental processes that promote the differentiated conformity called for by the distribution of culture inherent in a social structure composed of articulated statuses. The issues still to be examined all concern the ways culture actually affects what community members do. These include how cultural elements affect individuals who do not share them, how cultural elements are organized, including given priorities, and how individuals use cultural elements and products to pursue goals whether they are aware of doing so or not.

Chapter 9 discusses how those who understand themselves as ill find medical treatment. Understandings identifying the signs of illness are widely shared, as are beliefs holding that a wide variety of kinds of medical care are available in Mombasa and that getting some kind of treatment can be beneficial.

In trying to understand how culture provides guidance in getting care for the sick, an obvious hypothesis, based on the customary invocation of "shared beliefs and values," is that the sick use an intrinsically organized schema of shared understandings as a guide to dealing with their illness. It begins with their recognizing themselves to be sick, proceeds through understandings about what to do when sickness of the kind they attribute to themselves is present, and leads to visiting medical practitioners whom they understand to be able to deal with the causes of their illness as they understand them.

In fact, the connection between the understanding that one is sick and the decision about what kind of physician to visit is only rarely composed of intrinsically related cultural elements of the sort mentioned in the hypothesis. Contrary to the hypothesis and quite different from the sorts of schemata used by, say, chess players, the common schema for sick Swahili is most often extrinsically organized.

Patients are found to have only the most limited understandings about body functioning and illnesses, and although most know there are many sorts of medical practitioners in town, they have few understandings about the nature of differences among them. The basis for their decision about what treatment to seek is usually the advice they receive, so that this advice is the basis for the connection between the understanding that they are ill and the decision to consult one sort of practitioner rather than another.

Given the central part played by the advice, it follows that their expectations of the adviser, since that is what makes the advice useful, are central to their treatment-seeking schema. Since the expectations involved in the relations between people are elements of the statuses they occupy, for the great majority of community members who do not share most medical understandings, it is the status system that makes that part of the group's culture effective.

Chapter 9 describes the elaborate set of understandings shared among practitioners of what some group members refer to as "traditional Swahili" medicine. These understandings are shared among professional practitioners and also among what I discovered was only a small, but articulate, group of dedicated amateurs. It is part of Swahili culture since a number of group members share it with one another, but even its main outlines are unknown to more than three-quarters of the group. Nevertheless, it is a functioning part of the culture that affects most group members at some time—often many times—in that it affects what treatment they receive when they are ill.

This is a case, then, of cultural elements affecting those who do not share them. There is nothing unusual about this; much of what happens in all groups depends on members being affected by cultural elements they do not share. The way Swahili patients get treatment, however, offers an opportunity to examine the process whereby cultural elements can influence even those who do not share them. In this case, the vehicle for the effects of unshared culture is advice that is followed, and the acceptance of that advice is the result of the expectations in the role involving the patient and the adviser.

My investigation showed that the adviser was a parent for a substantial majority of those who lived with or very close to one or more parents and a spouse, neighbor, or kinsman for the remainder. The patients said that they followed the advice they got either because the advice giver "knows about illness" or because he or she had had a similar illness and reported being

helped by the practitioner recommended. On interviewing as many advisers as I could, I found that only about half themselves shared the basic understandings about body functioning and illness that would make an intrinsically organized illness-treating schema possible. The advisers, in many cases, were themselves advised in finding the practitioners they recommended; whole chains of advice expand the reach of the medical understandings to those who do not share them. This proves to be as true of understandings about and use of Western medicine as it is of traditional Swahili medicine.

For most patients, what is crucial to their getting the medical treatment they do are the expectations in the advisers' statuses and their part in their roles vis-à-vis the patients. As with Fernandez's (1965) ritual attenders who shared few understandings about the ritual they participated in with the specialists who arranged and staged the ritual, the patients' behavior is to be understood as a product of the understandings that connect statuses rather than directly as guided by the understandings concerning the immediate focus of the actual behavior.

As noted earlier, the distribution of understandings among statuses includes general expectations as well as specific ones. Patients told me, for example, that they followed their mothers' (and in a few cases, their fathers') advice about getting treatment because "my mother knows about these things," because "she is concerned about me and my health," and similar reasons. The patients "trust," "believe in," or (in a few cases) "obey" their adviser, and these views are rooted in general expectations about the adviser as part of the adviser's status as parent, spouse, kinsman, or neighbor. The adviser, in turn, has general expectations regarding whomever he or she found out about the therapist from. The intersection or connection of the general expectations in the different relationships (i.e., the adviser's direct or indirect trust in the therapist's ability and the patient's usually direct trust in the adviser) leads the patient to get treatment from a person whose relevant status is based on understandings shared with other therapists and with a few interested group members but often not with the patient.

It is the relationships among people guided by elements in their statuses, most especially general expectations, that lead the patient to seek and accept the treatment. If we think of social structure as the connections among statuses based in the mutual referring understandings that constitute those statuses, it is social structure, a product of culture,^[8] rather than the cultural elements concerning illness and treatment acting directly as a guide to behavior that accounts for what is observed in at least some group members' choice of therapy and therapists. Culture's elements, concerned with who trusts who as well as with who has understandings about the sources of illness or who is an acceptable therapist, are distributed among the statuses, and the distribution itself has a key part in the phenomenon of consulting a particular therapist.

Cultural Distribution and Social Structure

In chapter 10, the distribution of understandings among different statuses and the general expectations connecting those statuses are seen to provide a powerful resource for wives based in the unique character of the spouse relationship. This resource is vital to the wives sometimes getting what they want from their husbands even though the husbands have the undoubted power to refuse. This social structural effect, like that involved in the seeking of medical treatment, is seen to work even though some of the key understandings involved are not shared by all—or, in this case, any—of the actors involved.

The husbands say they give in to their wives' wishes because they do not want them to be "unhappy," and the wives say the husbands do it because of their "love" for them and because the husbands are "good." The source of the husbands' concerns about their wives' unhappiness and of their goodness is shown to be related to the fact that in all other relationships, men are blocked in receiving emotional support. The general expectations concerning emotional content in the husband's whole network of relationships makes the spouse relationship unique in this respect and thus gives the wife a powerful resource whether she admits it or not and regardless of her awareness of it.

Again, then, the distribution of culture is itself an important agent in influencing behavior. The effectiveness of the distribution of culture depends, of course, on conformity to the understandings involved. The understandings that people share can only be a basis for social life if the understandings concerned are used as actual guides to behavior. Since there are some understandings that apply to all or nearly all group members and some that apply only to those classified in particular statuses under limited conditions, there must be two different types of pressures for conformity. One of these must be parallel to the differential distribution of culture in order to produce the differentiated behavior appropriate to members of different statuses in various contexts. The other sort of pressure for conformity must be fairly uniform leading people to behave in somewhat similar, rather than differentiated, ways in certain domains and contexts where such similarity sometimes promotes group cohesiveness.

The Organization of Culture by Statuses

Chapter 11 brings together my findings and interpretations to summarize what has been found for the Mombasa Swahili and to propose a generally applicable status-based theory of culture's functioning. One of the theory's

main propositions concerns how culture's parts are related to one another, that is, how they are organized, to form an effective guide to action. One source of this cultural organization is "organizing understandings" that explicitly indicate the relationship between other understandings, as in "better safe than sorry."

The relationships among elements may be ones of relative importance as in the example just cited, but they may also concern sequence, dependence or independence, or other sorts of relations. Whichever sort of relationship may be involved, organization is an indispensable part of culture's functioning. Although explicit organizing understandings can contribute to the relations among cultural elements, such understandings are by no means the only or, necessarily, the most common or effective contributors. Organization may, and often does, result from understandings that have no intrinsic relations with the issues to be acted on. Put simply, the relationship between cultural elements is mediated by, even sometimes produced by, the understandings that guide the relationships between people.

Organization is vital because it directs the unavoidable choices that must be made among the understandings that might be useful guides in particular contexts and situations. This choice is often on the basis of actors' status-guided relationships to other people concerned with the action at issue rather than on the basis of intrinsic relationships among the usually numerous alternative understandings that are more or less applicable to that action. That is, actor X follows understanding A rather than understanding B not because A is more important, urgent, or virtuous than B but because the relationships with actor Y makes A more appropriate. Status-guided organization of this sort results from at least three somewhat different sorts of processes.

One is the type involved in the previously described spouse relationship. Stated generally, participation in one relationship is affected by the involvement of the participants in other relationships. This leads some understandings available in the relationship to be given precedence over others of similar sorts, as in the husbands withholding their power to refuse their wives' requests for finery and expensive ceremonies.

A related but different social process that serves to organize cultural elements is described in chapter 9. Here it is seen that many individuals decide what medical care to seek on the basis of the general expectation in a relationship—typically with the mother—that is not primarily medical in nature but that involves wide-scope trust. The understanding that leads to action is not one concerned specifically with illness and treatment but rather with the adviser's commitment to the patient's welfare. Similarly, the adviser's recommendation is rather often based on her trust of still a third person in a relationship with her which is not primarily medical. Fairly often, such links form a chain of nonmedical relationships that influence or determine decisions about medical problems despite few of the participants sharing even basic understandings about medical issues.

In this way, the effects of cultural elements shared among a few members of a group are "passed" from relationship to relationship and affect people who may be quite ignorant of the particular understandings that directly bear on, and may be crucial to dealing with, the problem or issue concerning them.

Somewhat similarly, the expectations in one relationship are transmitted to another through their effects on the person who is involved in them both. The person with dual involvement transmits the effects of the expectations in one relationship to his partner in the other with that partner's reaction depending not on the provenance of the expectations but, rather, on the latter's commitment to the relationship itself. The partner may transmit those effects to still another relationship, thus continuing the "chain reaction." This organization of elements involving bringing understandings into some kind of sequence need not be itself based on culture since it does not depend on those involved having any understanding of the overall interlocking network of expectations that actually produces the organization. A whole series of understandings can be linked together in this way, with the vital connection between them being their involvement in relationships with participants who have relationships with others "higher" in the chain rather than any intrinsic relations among the understandings themselves.

The main hypotheses in this book were developed through work in the Mombasa Swahili community, but they are proposed as being universally applicable. It is not part of the proposal that the processes found among the Swahili are the sole means for promoting culture's effectiveness. It is, however, a contention that these processes are found in all groups. The status-based organization of culture may be only one of a variety of solutions to the problem of culture's ability to function despite limited sharing within groups and, even, within statuses. Similarly, it may be that it can be shown that the nature of general expectations is less central to social life than it appears to be in Old Town and that the use of cultural models embedded in relationship terms is only one of myriad devices that teach understandings to adults while, at the same time, giving those understandings an emotional and evaluative weight that makes them harder to flout. It may even be that the importance of evaluation and its dual dependence on status is overstated and that other processes play a greater part in differential cultural conformity and the maintenance of social structure in other communities. Still, all of these processes are being presented as universal even though the data presented here are from the Mombasa Swahili community alone.

And now it is time to turn to those data. To put flesh on the conceptual bones and to provide a basis for assessing the claimed significance of the cultural processes described here, we begin with a consideration of the rise of the Mombasa Swahili community and its evolution over time.