106 Impasse and Resolution
ritual and belief. For Beattie, as we have seen, rituals and religious beliefs are expressive; for Horton they are to be explained as cognitive, rational attempts at explanation.5 Once again, we have an example where both scholars are excellent representatives of the functionalist theory at work in the social sciences despite a fundamental disagreement between them about the cognitive status of religious beliefs.
Functionalism can be viewed as the theory for explaining things in the social sciences. Most of the time the theory is never described in any detail but simply assumed as a valid model of explanation. This is especially true of scholars who have become popular in the study of religion. Mary Douglas's book, Purity And Danger, is a good example. The thesis of the book may be summed up as follows. Anomaly creates disturbances of a high level in undifferentiated societies. Such disturbances must be reduced or removed in order for the society to function adequately. Classification systems, taxonomies, and taboo function to reduce anomalies in a society and thus create solidarity by reducing disturbances created by anomalies. Clifford Geertz's elegant essays are clearly located within functionalism, as is Victor Turner's description of the function of "communitas." Festinger's explanation of religious movements based upon "cognitive dissonance" assumes the same functionalist model. All of these scholars have become popular resources for students interested in the study of religion. This is especially true for those who have recognized that the tradition of Otto, van der Leeuw and Eliade has reached a dead end. It is also true for many Biblical scholars who have discovered the importance of the "social context" for interpreting texts.
The theory is very persuasive. You need not really bother about why people hold religious beliefs. All you need do is recognize that the beliefs are relative to a cultural system. As Barnes puts it, "If we ask why an individual believes 'X' the usual answer in all cultures will be that he was told or taught 'X' by a trusted knowledge source. We can then ask why belief 'X' is present in the culture, and that is where functionalism makes claims to provide explanation."6 Indeed it does! The problem is that most scholars who use this model of explanation do not pause to critically reflect on the validity of their claims. Why, then, do they believe it? Well, they were told or taught the theory by a trusted knowledge source! How else do we explain the persistence of such a widespread theory of explanation which is seriously defective.
Functional Explanations of Religion    107
We can simplify the theory of functionalism as an explanation of religion in the following way: If i (e.g. religion) then z_, (e.g. anxiety reduction) where z is a functional requirement of x, (e.g. social maintenance) and y satisfies z. This formula was fully described in 1964 by Melford Spiro. His article, "Causes, Functions, and Cross-Cousin Marriage: An Essay In Anthropological Explanation" makes explicit the claim that functional explanations are causal explanations with an important difference.7 The difference can be described in the following way. Functional explanations are like causal explanations in that they also want to account for y (religion, for example) in causal terms. The difference, however, is that functional explanations explain y_ by reference to some condition z, in which z constitutes the contribution of y necessary to the maintenance of some social system x. Thus, the satisfaction of z is the function of y. Religion, for example, is not simply explained as functioning to satisfy social maintenance. Religion functions to reduce anxiety which constitutes the contribution of religion for satisfying a necessary need for the maintenance of a social system. The Kula ritual performed by Trobriand society is explained by reference to reduction of social tensions that must be satisfied for the maintenance of the social system.
The shift here is important. We explain a religion by showing how it satisfies a functional requirement of a social system. The functional requirement is usually expressed as some "need" in the social system. The use of functionalism in psychology focuses on "needs" in the personality system.
I shall not reproduce Spiro's functional explanation for the existence of cross-cousin marriage. I shall, instead, present a simplified model using religion as an example. I do not believe that this simplified model, distorts the claims made about functionalism as an explanation. The simplified version for an explanation of religion can be described as follows:
1.  x = a particular social system under a setting of kind c.
2.  y = a particular structural unit, (religion or ritual) which is a sufficient condition for satisfying z,
3.  z = a functional requirement (let us say family confidence and trust) which is a necessary condition for x, the social system.