The methodological approach which I have taken here, treats the statistical structure as a descriptive device and the normative structure as the relevant stock of actors' knowledge manipulated by them in the process of their decision making.  [...]  As far as the normative structure is concerned, what has been treated as problematic here is not the structure itself, but rather whether, why, and how it does or does not enter into individuals' decisions, or, in other words, what people do with it at the transactional level.  The problematic elements in my approach is thus neither the statistical nor the normative form of society, but the process of individual decisions which generate them.

from
Ladislav Holy (1979) 'Changing norms of inheritance among the Toka of Zambia.' in D. Riches, ed. The conceptualization and explanation of processes of social change [The Queens University papers in social anthropology, 3]. Belfast: Queens University, 103.

The invocation of a norm by actors in actual social transactions, and  the revalidation of the same norm through these social transactions, are two aspects of the dialectical relationship between actions and norms.  Any norm enters into an action only through being invoked and enacted in that action; at the same time, the action in which the norm has been enacted affects this norm in that it revalidates it and thus makes it invocable and enactable in a future action.

from
Ladislav Holy and Milan Stuchlik (1983). Actions, norms and representations: foundations of anthropological enquiry.  Cambridge: Cambridge University Press, p.121

The anthropologist's acknowledgement of the role of actors' notions in shaping social reality does not mean that his explanation must resemble the actors' own model or be a duplication of it. In fact, it will never be either for the simple reason that the anthropologist's and the actor's interests in the same social reality fundamentally differ: the actors' interests are always practical, the anthropologist's interests are theoretical. Guided by his theoretical interest in the social reality he studies, the anthropologist will always ask of it questions in his analysis which are not asked and cannot be asked by the actors. He will seek answers to problems which are not perceived and cannot be perceived as problems by the actors. In consequence, the anthropologist's explanatory model has to be by definition different from any kind of model the actors have. The anthropologist's explanation is fully legitimate so long as it does not alter the meaning which the phenomena explained have for their own creators, i.e. so long as the actors' meanings are not replaced by the anthropologist's meanings. In our references to existing analyses at various points in the essay we have tried to indicate some of the ways in which such a substitution of meaning occurs. When actors' meanings are replaced in the course of analysis and explanation, the anthropologist is not explaining social reality as it exists in the only meaningfully possible sense, but through his explanation creating it. Since social reality exists only as a meaningful reality, it is through creating meaning that social reality itself is created.