The relationships that are developed with friends in the field are inevitably based on reciprocity and compatible learning, and through this it is easy, and perhaps common, to launch 'informants' in intellectual careers as analysts of their own society. In this capacity they can naturally provide a number of excellent insights – but these insights can not be used as data on the contents of their cultural tradition. And once this process is launched, it becomes impossible for the anthropologist to tell which items are inherent in the tradition he is studying, and which represnt the feedback of his own activity. Working as I did with an essentially uncontacted population, it seemed particularly valuable to avoid such adulteration of the material. [...]

Careful observation of this procedure led to the empirical discovery that the Baktaman have no exegetical tradition. I am struck by the lack of factual comparative evidence in the anthropological literature on the presence and degree of development of native exegetical tradition and praxis; and without the careful restraint I practiced, I might not have recorded its absence. As social relations in the field deepened, it would not have been difficult to obtain native help in my efforts to understand: to make them systematize and translate verbally in response to my need for system and verbal codification. My strong suspicion is that the bodies of native explanation that we find in anthropological literature are often created as an artefact of the anthropologist's activity.

Barth, Fredrik. 1975. Ritual and Knowledge among the Baktaman of New Guinea. Oslo: Universitets­forlaget, p.225-226.