Early Austronesian Social Organization: The Evidence of Language [and Comments and Reply]

Abstract

Published views on early Austronesian social organization can be characterized broadly as representing one of two mutually irreconcilable positions: one that recognizes descent groups and another that does not. The former position-in a highly specific form that I call the "prescriptive-alliance hypothesis"-has been adopted by most Dutch scholars concerned with the culture history of Indonesia and was independently reached by Levi-Strauss in the context of a study with wider implications. A distinct variant of it has been defended by Goodenough. The opposed position-which I call the "bilateral hypothesis"-was apparently first expressed in modern form by Kroeber with reference ot the Philippines and later by Loeb with reference to Indonesia. In more recent years it has been defended most cogently by Murdock. The present paper attempts to resolve this conflict of inferences on the basis of comparative linguistic evidence, following a set of methodological procedures similar to those advocated by Dyen and Aberle. At least ten independent linguistic observations or groups of observations are found to be anomalous within the framework of the bilateral hypothesis but become intelligible on the assumption that early Austronesian society was (1) based on prescriptive alliance, (2) divided into "male" and "female" moietes, and (3) transected by a second dual division which gave rise to a quadripartite social organization (presumably a four-class marriage system). It is suggested that Murdock's system of probable transitions between types of social structure is erroneous either in principle or in application to the particular case of Austronesian and that descent groups have been lost within the past two to three millennia over a generally continous area in western Indonesia and the Philippines.