Who is a What?

For statuses to work as they do, participants in social relationships must share with their partners in interaction understandings of what categories each is assigned. As far as simplex relationships are concerned, recognition of own status and that of the other seems to depend on little beyond an understanding of what the interaction is about. If one had never been on a bus,^[4] it might be difficult to understand what being a passenger involves as well as what a "bus driver" is, what he expects, what can be expected of him, and the range of circumstances in which the driver-passenger role is salient. Even in this unlikely case, however, one would quickly learn or stop trying to ride buses. The same is true, mutatis mutandis, for many other simplex statuses.

Even for multiplex statuses, there is so much agreement about categorizing those in broad-scale statuses that it is difficult to get people to discuss the issue. An extract from my field notes on a discussion concerning how one knows who is assigned to a category indicates something of what is encountered when such inquiry is approached directly. The informant, "BR," is a woman in her late fifties whose husband is dead and who lives with her two sons, one seventeen and one twenty-three.

MS: (BR has just given me an example of what a friend is by discussing a friend of her oldest son. The son's friend is Hamid.) How do you recognize that Hamid is your son's friend?

BR: I know he is my son's friend because my son has known him for a long time and they go around together.

MS: But how is he different from other young men whom your son has also known for a long time?

BR: They like each other. They are friends.

MS: So a friend is anyone you like and you go around with?

BR: Yes, that is a friend. (If he goes away?) Yes, if he goes away [as to another city or country], he can still be a friend if they continue to get along and remember each other.

MS: But, so I can be sure I understand, when does someone become a friend?

BR: Don't you have friends in America? It's like how you know your neighbor or anybody. You know them.

However limited cultural sharing may be, the idea that one may be in doubt as to who is a friend, an enemy, a brother, or a neighbor is seen as ridiculous. It may be that a person identifies another as a friend or a neighbor but is not similarly identified by the other who, instead, classifies the first as an enemy or vague acquaintance. As far as their direct interaction is concerned, this need make no difference so long as each behaves according to the expectations held by the other as concerns whatever interaction is taking place.

It may be that one acts toward the other as friend and the other acts toward the first as "person who thinks he is a friend." The overlap of the expectations in the two categories may well be sufficient to allow direct interaction of some kinds to proceed. Difficulties would emerge only when one of those involved followed an expectation (e.g., getting help) that is not consistent with the other's classification. Even when direct interaction is proceeding smoothly, the disagreement in classification may affect other relationships (where participants may know of the unshared classification) in ways that shared mutual classification would not.

I have been told by an older man that a younger was an mshenzi (uncultivated, uncouth person; see below), but when I saw the two interacting in the limited ways they did, as when entering a mosque at the same time, they acted toward one another as others do who share the community member status. It is clear, however, that the older man's "true" (or, at least, unfavorable) identification of the other affects his behavior in some situations since he did tell me the young man was an mshenzi. In fact, neither identification excludes the other, and it is a matter decided by his salience understandings as to which status, or combination of statuses, guides the older man's behavior toward the younger.

So long as the older man treated the younger as "community member," the young man responded to the expectations in that status and treated his elder, in turn, as "community member." The identifying understandings each used resulted from the mutually contingent status assignment each made of the other so that the interaction proceeded. It is likely it would not have continued as it was when the two met at the mosque had the young man wanted

to marry the daughter of the elder, since that would involve expectations beyond the "community member" status, and these would probably touch on the elder's expectations associated with the "mshenzi" status.

From the point of view of how culture works, it appears that agreement on mutual classification is not difficult to achieve in face-to-face relations and depends as much on mutual adjustment of expectations as it does on a preexisting set of shared understandings. Multiple identifications and assignment to more than one status at a time provide flexibility in different settings according to their different expectations without necessarily obstructing interaction in any of them. There is an effective limitation on conflict arising from unshared mutual classification. This is that those who cannot agree on mutually acceptable categories for one another most commonly cease interaction.

In simplex relationships, the basic identifications involved are nearly unavoidable and leave little room for error. Bus drivers and passengers either place one another in those statuses or interaction concerned with bus riding is impossible.^[5] In multiplex relationships, mutual identification is also vital, but, as with the example above in which one person classified the other as friend while being classified himself as enemy, the identifications may be only partially shared. A number of Swahili men, mainly former administrative officers,^[6] have told me that in the colonial days, the British officials who came to the coast from Nairobi did not "know how to act." These outsiders, unsocialized by the Swahili and Arabs of the coast, treated all "Africans" the same—to the Swahili, an obvious failure to make important distinctions and one that made the upcountry colonials undesirable partners in interaction.

As Holland (1987a ) discovered in her important study of American college students, the ability to differentiate among statuses depends in part on the interest and involvement of the actors in interaction with one another. In what she calls "the romantic sphere" (i.e., the relations of male and female students in dating and such) she found substantial agreement among the students concerning the elaborate classifications of their associates, but in the "academic sphere," where classification was according to subjects studied and similar matters, there is much less agreement (ibid., 240–242).

This finding is important here. It may be that identifying understandings even for the statuses active in multiplex relationships are not completely shared. That they should be fully and generally shared seems unlikely given the generally limited nature of cultural sharing. Since, however, these statuses always involve crucial areas of the lives of those involved, it is likely that the sharing of mutual identification is very substantial. This is made more so by the operation of the mutual adjustment that is the heart of double contingency (see below) in the frequent interaction characteristic of such relations. This is probably the reason Holland's results show that those in high-frequency interaction, common to multiplex relations, come to agree on mutual assignment.

It is important to note that the behavioral importance of mutual assignment in large part resides in the expectations that are part of that assignment. Some expectations are quite specific, and behavior is easily assessed as having met them or not. Other expectations are broader and more diffuse, and establishing whether these have been met involves substantial interpretation. In multiplex relationships, specific expectations and salience understandings of the "passenger gives money/driver issues ticket" sort are usually distinctly secondary. The general tone and significance each participant assigns the behavior of the other matters more in these relationships, and there is a great deal of room for variation in what is acceptable.