"Tokens" and "Guides"

This would explain the disparity between the presence of important differences between the members of different generations as seen in various be-

haviors such as dress and demeanor and the absence of evidence of these differences in formal interviews. It suggests that shared understandings need not be related to behavior only or mainly as guides for that behavior. The responses in formal interviews can serve as symbols whose exact reference is relatively unimportant but whose proper production (i.e., when to produce them and who to give them to) is crucial. This is a point that the ethnomethodologists in sociology have examined in many important studies (e.g., Cicourel 1987).

It is not that these—or most—symbolic expressions of understandings, that is, tokens, have no relation to behavior. Rather, it is that the relationship is not necessarily what it might appear to be. Saying that one is responsible for generational tensions is not the statement of an understanding that guides behavior in intergenerational relations. Instead, it is the production of a symbol seen as important in their relationship with the interviewers, the anthropologist who employs the interviewers, and their community as a whole. In effect, they are asserting that they share with their fellow group members vital understandings about aspects of group life, and this assertion is by means of tokens—in this case, statements about behavior—given to those with whom they are speaking.

In fact, the use of tokens is by no means limited to responses to interview questions. One of the consequences of successfully passing tokens is the affirmation of various memberships and commitments. Swahili of all ages, for example, exclaim Hamdulila (God be praised) when things go well and Allahuakbar (God is great) when surprised, and begin most undertakings by asserting Besmilla (In God's name). This asserts their unity with other Muslims and separates them from Christians and others, just as their dialect of the Swahili language differentiates them from Swahili speakers belonging to other groups. Carol Eastman has argued that language is "culture loaded" (1979) and that it serves to establish social identity (1984, 1985). This applies to the Swahili, with whom she has worked, as well as to speakers of other languages.

More than asserting identities and commitments, however, exchanging tokens also provides confidence that those exchanging them share important understandings. The parents who say they are responsible for their children's behavior are asserting the existence of common ground with those children and anyone who sides with them, and, mutatis mutandis, the children's assertion of responsibility for themselves does the same with their parents and their allies.

The distinction between cultural elements serving as tokens and as guides is useful in understanding how culture works despite the quite incomplete sharing of many of its constituent parts. Malinowski, Fernandez, and, probably, many others saw the importance of this distinction in accounting for some aspects of culture's functioning. Surely, although Reisman is an unusually accomplished and sensitive field-worker, he is not the only one who has

noted the importance of tokens and their frequency of use in particular areas of life.

Swahili men, to take another example, wear kanzu and kofia as other Muslims do. But Swahili men, and no others, wear white kanzu embroidered with small beige designs on the placket and kofia that are white with white embroidery. These are tokens presented to all who see them which will be understood by those who matter most as what they are: assertions of group membership and acceptance of group understandings.

Some tokens, then, are purely communicative symbols as well as statements of understandings that do not actually reflect guides for behavior but, rather, serve to assert the existence of sharing. It is doubtlessly also true that some guides are also tokens. Among the Swahili, for example, proper performance of the daily prayers is a crucial token that community members give one another every time they pray publicly.^[4] The understandings that lie behind this token are, so far as I can tell, nearly completely shared.

Every Swahili more or less fully shares the Sunni view of how to pray and the meaning of the movements in prayer taught him or her as a child. Using informal evidence, the understandings about the meanings, unlike the understandings about the Bwiti cult ritual, are nearly universally shared. These shared understandings about prayer actually guide people in the way they pray. Nevertheless, the way praying is done serves as a token of group membership, joint and shared belief, and a raft of common values and beliefs about the supernatural, humanity, and the world. It is not sufficient to pray properly to be recognized as a Swahili—various Muslims of Indian origin and from other African groups also do it—but it is a necessary token.

Understandings need not be divorced from the activities they concern in order to serve as the basis for tokens. They may, in other words, be both guides and the basis for tokens. The willingness of group members to use and to accept tokens, in fact, probably rests in considerable part on the fact that some tokens are also guides. Since they sometimes are, it is not groundless to take symbolic expressions of agreement, unity, and solidarity as at least possibly reflecting the true (i.e., active and effective) condition. The understandings that are not, in fact, guides for the behavior they address might be; others are. The fact that Swahili prayer behavior actually symbolizes understandings that are guides for other behaviors and beliefs, together with other behaviors that both assert sharing and actually guide activity, contributes to the appeal of making such statements as those about intergenerational relations.

Tokens may be consciously intended as such or not, but they always serve to indicate the presence of particular understandings. Sometimes tokens can be recognized because, as with the responses to the survey interviews, they are clearly not guides, but they can also be recognized by finding out what significance others attach to what people say, do, and otherwise communicate.

The basic distinction between "guides" and "tokens" is a functional one.

Guides are understandings applied as determinants of behavior in the areas indicated by the statements. The Swahili understanding that the left hand is not properly used in eating is readily expressed and often heard being directed to small children and anthropologists. Observing the Swahili eat, one sees the left hand kept out of action even for rather difficult one-handed tasks such as breaking a piece off the flat, pancake-shaped breads the Swahili often eat. Seeing the eating, one concludes that the expressed understandings about using the right hand and avoiding the left hand are guides. They also serve as tokens, of course, but they are quite redundant ones in most contexts where group membership and the sharing of understandings about eating are symbolized by the way the people sit, the food they prepare and eat, and so on.

Regarding the behavior that was the focus of much of the discussion of intergenerational relations in the interviews, young people report in informal conversations that there is nothing wrong with how they act in public. Difficulty, they say, arises from the fact that older people use inappropriate standards for judging them. Taking their statements as representing understandings concerning their evaluation of their own behavior, these understandings are clearly guides. This view of the understandings is confirmed by the fact that they continue to behave in the ways they say are unobjectionable without changing what they do to take account of the objections of the older people.

Guides, to put this generally, are understandings whose effects can always be seen in behavior beyond (in addition to) statements about what is or should be. Tokens may or may not be linked to observable activity beyond communication. The existence and use of tokens is an important support for social life and for those cultural processes, to be examined in subsequent chapters, that depend on social life.

Recognizing the possibility that understandings can be shared but connected to behavior only as tokens rather than as guides, or guides that are also tokens, appears to have analytic usefulness. It suggests an explanation for phenomena such as the two sets of data concerning the Swahili generation gap. As will be seen below, it also seems to make easier the recognition of a process that plays an important role in culture's ability, despite its quite incomplete sharing, to serve as effectively as it does as the basis for social life.