There can be no question that the period from the ninth to eleventh centuries in westem Europe was one of political upheaval and change for the aristocracy. Charlemagne's empire was invaded, fought over, divided into new kingdoms and principalities. Fief-holding, vassalage, and castles first became widespread. Even the sorts of men who wielded power changed as new lineages first of counts and then of castellans appeared and married into previously established lines1. This political change, it

more »

... generally agreed, was accompanied by some sort of change in the family structure of the aristocracy, but there has been a good deal of debate over exactly what this change entailed. In this paper, I shall reexamine the question of noble family structure in this period, trying first to define some of the parameters of the discussion and then making suggestions on the nature of the changes in family consciousness, suggestions quite different from the conclusions many have drawn in the last twenty-five years. I shall do so using concrete examples drawn from three different lineages or family groups. These are the Carolingians, whose power and authority was unquestioned from the eighth Century on, the Bosonids, who tried with greater or lesser success to become kings and emperors in the ninth and tenth centuries, and finally the dukes of Septimania and Aquitaine, a group that is one of the few well-documented non-royal lineages that can be traced throughout the ninth and early tenth centuries and which has been taken as a stirps classicus of west Frankish society2. Recent studies of medieval family consciousness, usually relying on the classic work of Karl Schmid, have drawn the conclusion that in the Carolingian age family structure was quite diffuse, involving a large, horizontally-organized group of cousins, both agnates and cognates, all alivc at the same time (a Sippe in German), and that only after the year 1000 did a verticallyorganized, patrilineal form of family structure (Geschlecht) become common. This »Schmidthesis« has grown beyond the rathcr modcst suggestions of Schmid himself to become in the