Social constraints on the generation of culturalrepresentations

One of the few anthropologists who has paid careful attention to the role of memory in his research on cultural transmission is Fredrik Barth in his classic study (1975) of the ritual and knowledge of the Baktaman of central New Guinea. We shall argue in the next section that the seven degrees of Baktaman male initiation demonstrate how performances of religious rituals exploit many of the variables that experimental psycholo- gists have shown make for more accurate memories of specific episodes. ¹² Before we pursue that line of argument, though, we shall first examine

-64-

Barth's own discussion of the transmission of ritual materials among the Baktaman.

The Baktaman are one of many Mountain Ok groups who reside in the central highlands of New Guinea. They numbered 183 (with six deaths and three births) at the time of Barth's fieldwork in 1968. (Barth returned to the region for three months in 1981–1982. ) All of the evidence indi- cates that prior to 1968 the Baktaman had only two fleeting contacts, in 1927 and 1964, with anyone outside related Ok groups occupying the lands within a few dozen miles or so of their own territory. Barth de- scribes the Baktaman as “persons entirely unacquainted with Man in any other form than themselves” (1975, p. 6).

Prior to “colonial pacification” in 1964, relations among these Ok groups were rocky. Although these groups are close enough to intermarry, to recognize clan relations across groups, and to invite one another to ob- serve religious rituals, suspicions of sorcery and comparable offenses led to frequent wars and raids. Barth's research indicates that one-third of Baktaman deaths over the previous two decades had resulted from such violence.

The system of male initiation among the Baktaman fulfills all of the conditions for an ideal test case outlined on pages 55–56 above. In addition, this system introduces a few additional barriers to ready trans- mission that we have not yet discussed. The Baktaman are a small, non- literate, isolated group of hunters and subsistence farmers. They move within their region every few years as their gardens exhaust the nearby land. Males go through seven degrees of (non-repeated) initiation in their lives. (A cohort of men — most of whom were in their thirties — had yet to undergo the seventh degree initiation when Barth arrived. ) Because entire age cohorts undergo these rituals together, they occur quite infrequently – approximately once every ten years.

All Ok have additional opportunities to observe what are, at least, roughly comparable initiations among their neighbors. Barth estimates that seniors have four or five such opportunities between performances in their own group. This is roughly once every two to two-and-a-half years (Barth, 1987, p. 25). However, since the number of neighboring groups is small, since unhappy relations sometimes preclude invitations, since the invitations involve only parts of the performances, since these other groups perform these rites no more frequently than the Baktaman do, and since other groups' rites differ from those of the Baktaman on

-65-

The rituals are long, ranging from a few days to a few weeks. Perhaps posing another barrier to easy recollection, earlier initiation rituals in the series are sometimes deceptive. Later rituals often reveal how earlier rituals were misleading on one point or another. On at least one count, the deception is double layered (across three degrees of initiation).

All of the rituals include parts kept secret from non-initiates, i.e., from females and junior males. ¹³ Baktaman expectations about secrecy are extraordinary. Senior males threaten initiates with death if they violate ritual secrecy. Not only are initiates prohibited from discussing these secrets with non-initiates; most of the time they are also prohibited from discussing them among themselves. Incredibly, they even seem to exhibit “a wariness and vagueness in thinking about them” (1975, p. 221). As far as Barth could tell, initiates respected the prohibitions on overt behaviors, at least. (One cult leader was so concerned with secrecy that he hid his clan's sacred relics in the forest, and when he died unexpectedly, these articles were lost. )

Barth maintains that secrecy is inimical to ready transmission of the re- ligious system's symbols in another way. The impact of these demands for secrecy is so substantial that it undermines any hope of forging widespread logical coherence among Baktaman beliefs. Barth argues at length that what he calls “analogic coding” grounds whatever order inheres in the cultural knowledge of the Baktaman (1975, pp. 207–231). “The medium is one of metaphor, as in the manipulation of sacred concrete objects and ritual acts to generate statements about fertility, dependence, etc. ” (Barth, 1987, p. 75).

What exactly is analogic coding? First of all, it is not what Barth refers to as “digital coding. ” Presumably, digital coding is where non-linguistic public representations are organized into systems manifesting properties something like a genuine code, with rules of correspondence and univocal connections between items. The Baktaman rites do exhibit some com- positional relations, since the Baktaman reuse a few symbolic objects and activities as components in larger symbolic configurations. However, whatever productivity they exhibit is not based on any systematicity ¹⁴ nor on any “logical closure” nor on some “limited set of alternatives” (Barth, 1975, p. 208; also see p. 229). Each setting in which the Baktaman reuse a ritual object involves what are otherwise undisclosed symbolic

-66-

Each symbol turns on “a very simple non-verbal metaphor sketchily exploited” yielding “complex harmonies. ” Barth (1987, p. 68) states that these non-verbal metaphors “somehow 'feel' compelling and right, but they are based on fortuitous analogies. ” Apparently, the underlying “metaphors” are “non-verbal, ” because the symbols are non-linguistic concrete objects (i.e., lasting, non-linguistic, public representations) and because the Baktaman are either unwilling to articulate these symbolic relations or incapable of doing so or both. ¹⁵

Barth explicates these metaphors and idioms in terms of general res- onances among the concrete symbols that figure in Baktaman myth and ritual from one symbolic context to the next. He identifies broad themes ¹⁶ – relations with the ancestors, fertility, security and welfare, etc. — with which various concrete symbols are associated in particular Baktaman rites. He maintains that it is these underlying thematic connections among concrete symbols across different contexts that constitute the knowledge Baktaman initiations present. He states “it is the themes of the rituals, not their sources of metaphor, that are explicated and ordered in the rites” (1975, p. 211).

Barth (1987, p. 76) suggests that participants in such a tradition possess this core knowledge intuitively. The anthropologist cannot reduce such knowledge to unambiguous propositional form, but this does not entail that either its contents or its effects are utterly random. “I am pleading for the recognition of positive achievements deriving from the potentialities of particular communicative forms… The medium in which the knowl- edge is cast allows other and rich forms of understanding, and directs its practitioners in large part to use them” (Barth, 1987, p. 76). Barth is fairly clear about those other forms of understanding analogic coding facilitates. He states (1987, p. 79) that these initiations transform “a group of young persons into men” who possess “a general area of common sensibilities and intuitions” and “a range of understandings sufficient so its members can be moved by the same symbols and thoughts. ” The initiations instill in the initiate distinctive cognitive dispositions and sensibilities concerning self, cohort, society, and Nature.

Barth (1975, p. 229) insists that “an analogic code must… be under- stood in the context of its praxis, ” noting that secrecy and the complete absence of exegesis and texts are the pivotal features of Baktaman praxis that have shaped this code's form and the transmission of Baktaman

-67-

The total corpus of Baktaman knowledge is stored in 183 Baktaman minds, aided only by a modest assemblage of cryptic concrete symbols (the meanings of which depend on the associations built up around them in the consciousness of a few seniors) and by limited, suspicious communication with the members of a few surrounding communities. (1975, p. 255)

Barth and his informants agree that it is the attempts by the seniors responsible for staging these initiations to recall past performances that are the primary influence on the shape of the next one (1987, p. 26). In Cosmologies in the Making (1987), Barth concentrates specifically on this process of recall and transmission of cultural knowledge in the various Ok groups, including the Baktaman.

Although he characterizes his specific approach as “generative” there, his project also fits into a more general framework, which Barth endorses, that can also be fairly described as “epidemiological. ” Barth (1987, pp. 83–84) finds the factors responsible for producing divergences in individuals' experiences and thoughts far more interesting than depicting “collective representations” (about which Barth expresses skepticism). He (1987, p. 28) too insists that “every person's mind is full of represen- tations of cultural objects, which are handled by mental processes and in due course give shape to the person's acts. ” For Barth, like Sperber, the principal task is to delineate the causal variables that shape the distribu- tions of cultural representations. He states:

My focus is on… this aggregate tradition of knowledge: the (variety of) ideas it contains, and how they are expressed; the pattern of their distribution, within communities and between communities; the processes of (re) production… and how they may explain its content and pattern of distribution; thus, the processes of creativity, transmission, and change. (1987, p. 1, emphasis added; see too p. 74)

Barth's project is generative in orientation, because he focuses on the creative processes that underwrite these (re) productions of cultural ma- terials and the forces that constrain them. Although recall may inform it, the overall process in question is just as much one of producing an initi- ation as it is one of reproducing a copy of an earlier performance of that initiation again.

-69-

Barth (1987, p. 84) advocates the study of “a living tradition of knowl- edge — not… a set of abstract ideas enshrined in collective representa- tions. ” Thus, he aims to discover “within a tradition of knowledge, the patterns of variation and thereby the underlying processes of thought, in- novation, and stimulus at work within it” (1987, pp. 19–20). Barth holds that the “basic problem is to find the most fertile way to conceptualize the locus and mechanism of incremental change, ” i.e., to identify “the mechanisms involved in such creativity” (1987, p. 27). All of this may look very much like Sperber's concern with specifying the micro-processes at the psychological level that affect the distributions of cultural representa- tions, but Barth locates the principal mechanism of incremental change at a different level of analysis. His chief concern is with the influence of Ok social arrangements on ritual experts' activities and mental states. Barth confines his psychological comments to but a few speculative observations (1987, pp. 30–31, 38, and 73).Barth (1987, p. 78) highlights four salient aspects of Ok social organi- zation:

-70-

The prevailing affirmation throughout Ok communities is that the an- cestors specified the details of male initiations. It seems impossible that without written records — perhaps even with them! — ritual experts could remember every relevant detail of a performance that lasted for days or even weeks and occurred ten years before. Among anthropologists who have studied the Mountain Ok, it is uncontroversial that they do not. During his initial fieldwork among the Baktaman, Barth (1975, pp. 239–40) detects nine examples of possible ritual innovations for which he has some evidence. Barth is not at all sure, though, about how aware Ok ritual experts are of the amount of innovation — beyond overt bor- rowing from another Ok group — each new performance involves. At least some of the time, it is clear that Ok ritual experts can improvise or guess. (See Barth, 1987, pp. 29 and 31. ) Barth notes that the conditions of transmission in these cases require that they do so.

A pivotal question is how much latitude ritual experts have when they face these periodic opportunities for cultural innovation. Barth (1987, pp. 55, 79, and 86) generally seems to see most of the centripetal forces in this process arising from Ok social arrangements. By contrast, what centrifugal forces arise come from what Barth takes to be less constrained aspects of the psychology of creativity, which Ok demands for secrecy only exacerbate. He says, for example, that his model envisions “an interplay of (largely divergent) processes of individual creativity and modification and (largely convergent) cross-influence and borrowing” (1987, p. 80; see too pp. 55, 79, and 86).

In addition to the comparatively conservative effects of borrowing, Barth identifies other social conditions, which, he believes, impose re- strictions on the ritual experts' freedom. Perhaps the most important is simply the expectations of various segments of the audiences to whom they must play and with whom they may consult in advance of the perfor- mance. The relevant audience segments include other experienced senior males and other ritual experts, who have their own recollections and con- ceptions about how things should go. Barth (1987, p. 29) suggests that decorum and collegiality usually insure that the resulting performances are “canalized but not controlled by the sanctions of each respective audience. ”

Barth (1987, pp. 62–64) also proposes that such variables as cohort size, the optional character of some higher degrees of initiation among some Ok groups, and the number of community-endorsed ritual ex- perts on hand can incline presiding ritual experts to explore some sym- bolic environs rather than others. Ok social arrangements and praxis also

-71-

Although the conditions of cultural transmission among the Ok yield “a movable, impressionable and everchanging inheritance, only approxi- mately shared in the group” (Barth, 1987, p. 79), Barth devotes most of his discussion to clarifying social forces that contribute to increased stability in Ok initiations. Like Barth and unlike Sperber, we too are identifying forces that tend to constrain variation in the transmission of religious rituals. On the other hand, like Sperber and unlike Barth,

-72-

the forces on which we are concentrating are psychological rather than social.