Data on the Lue of Chiengkham, Thailand, illustrate a contention that much of cultural anthropology consists of (1) reporting the folk predicates of folk ethnic identification labels, (2) assuming that all predicates are properly ascribable to such labels, and (3) looking for human populations to which the labels can be applied. In this paper, participant observation is criticized as a technique for discovering native categories since ethnic labeling appears to be motivated rather than self-explanatory. The discussion of the motivated nature of labeling appears to ascribe the factor with a retrospective, rather than an inductive use of traits. (Author/AF)