Barth was an extremely distinguished anthropologist, who did ethnographic fieldwork all over the world. This is an extremely unusual venture into anthropological theory that grew directly out of trying to make sense of a very puzzling aspect of that fieldwork. In 1969, Barth lived among the Baktaman, "a community of 185 inhabitants" in highland New Guinea, where he documented, and participated in, an extremely elaborate series of (male) initiation rituals and their associated mythology. In 1981--1982, he revisited Baktaman, but also a number of other communities which also were part of the same "Mountain Ok" group of peoples, "a total population of c. 15,000 individuals, of six closely related language communities, occupying c. 10,000 $\mathrm{km}^2$ of upland forests and mountains" (p. 2). He also drew on the writings of about a dozen other anthropologists who had in the meanwhile done work among the Mountain Ok. His goal was to try to make sense of why all these communities had different rituals and myths.
What Barth frankly confesses is that he has a very hard time making sense of the differences in ritual across the research sites. The usual tactics of anthropologists, and almost all other social scientists back to the Enlightenment, would be to explain or interpret the differences in rituals and myths in terms of differences in these communities' ways of life, their ecological niches or internal power relations. Maybe cultural content "reflects" social structure, maybe the former "expresses" the latter, maybe culture is "adaptive" or "functional", maybe they are moments in a complex dialectical relationship that resists reductive unidirectional and monocausal accounts, but even that's some kind of explanatory connection (supposedly). Barth tries valiantly, only to confess defeat: the differences in life-ways are small, their correlation with ritual and mythic differences nugatory and utterly resistant to interpretation. This being the early 1980s, he gives a Levi-Straussian working out of binary oppositions a try, but that too goes nowhere. (I suspect his heart wasn't in it.)
In something close to despair, Barth is driven to considering the cognitive processes at work within the community's ritual specialists. The initiatees are groups of men in the same age-set; the rituals are conducted infrequently, for some of the most important only once every ten years; in between, the full ritual and myth may be known only to one or two specialists. Those who have been through the rituals don't discuss them with the un-initiated, and rarely with each other. (The ritual specialists at Barth's new field-site were willing to talk about their rites and myths with him, because they recognized his earlier initiations from the Baktaman as valid.) Sometimes the initiator dies between rituals and parts of the rite or myth are simply lost. It's very important that the rituals be emotionally compelling to the initiates, and they involve stress, physical discomfort, and not having very clear memories of exactly what happened (while being deeply moved). All of these communities are illiterate, and have no other means of recording rituals.
Under these circumstances, then, the ritualists almost have to make stuff up each time they prepare and enact the ritual. In doing so, they will draw on their own memories, symbolism and analogies that will be familiar and moving to their fellows, and perhaps ideas from neighboring communities. (Remember that these communities [mostly] recognize each others' initiations.) This will naturally introduce variation over time and over communities, even without any of the usual just-so-story considerations about expressing or reflecting social structure. It also makes it very difficult to regard the myths and rituals collected on any one occasion as somehow expressing a collective world-view, as opposed to the vision of an individual thinker.
Barth does not, alas, actually articulate a generative model for Mountain Ok initiation rituals. (For all his many virtues, he was no V. I. Propp.) But he points in the direction of the kind of phenomena a generative model ought to explain, and the kind of mechanisms it ought to invoke. Moreover, in doing so his writing shows equal respect to his fellow anthropologists and to his research subjects. Even if you don't have any special interest in the native cultures of New Guinea, anthropological theory, etc., there's simply a lot to ponder here about human creativity, human myths, and the ties between the two.