52 SOCIAL SCIENTIST be whether it is in terms of 'political will' or in terms of the class character of the state (Herring, 1983). The second framework that has been used to analyse agrarian systems can, for want of a better term, be referred to as the liberal point of view. In this view it is the distribution of land that is the key to understanding agrarian systems. The large farmers by virtue of the fact that they own more land have greater access to inputs and greater control over the markets (e.g., Griffen, 1974). In such a framework it is the redistribution of land which is considered the primary measure of agrarian reform and even, as in the case ofLipton (1974), the only measure. Such a framework necessarily underplays the role of indebtedness and is thus not suited for our purpose. The Marxian framework is perhaps the most developed for the analysis of agrarian reform as it can take into account the movement from a system in which extra economic factors are dominant to a system donxi-nated by economic factors. In addition, the system has social and political considerations built into it. In practice, however, it is assumed that land is the major factor through which control is exercised in all agrarian systems. The model of a feudal system consisting of land owned by the landlord and cultivated by the tenants dominates the literature though it is now being increasingly recognised that it has its limitations as c^n. be seen in this summary by Pcarse (1983, p. 45) : . . . the most frequent pattern is for a single large landowner to le^se small parcels of land to numerous tenants who probably possess little or no land or other means of production of their own. In India, the situation is more complex. First, a substantial (though disputed) proportion of share croppers own some land of their own and second, thfere exists a wide variation in landowner-tenant relationship, from the situation . . . where the position of tenants is little different froip that of indentured labour, a situation more characteristic of eastern India', to 'the case of comparatively well-to-do landowning peasants teasing in Itod frotti smalt scale owners. In such a 'view indebtedness is feeen as only accentuating or retardiq& relatfoilships that are determined by the nature of land ownership. This perception cannot t^ke into account systems in which all the land was held1 by the feudalE sovereign and control over the village level agrarian systeih'was dxercised thr6ugh factors other than land, such as indebtedness. And such systems did e^cist in several parts of India including southern Karnataka and parts of Assam, where there is very little evidence to suggest tMat tenancy Was a dominant relationship even in the past. Two Extremesf In order to gain a clearer understanding of the role of indebtedness