PEASANTS, POLITICS AND HISTORIOGRAPHY 59 evidence of "sporadic" and "spontaneous" revolts does not- warrant the conclusion that "peasant politics" is autonomous. There is always "the possibility of manipulation, incitement and confusion" created by the interventions of dominant classes. In the evolving state structure of the late colonial period in India, if any group acquires a measure of autonomy, it is, he says, the landlords. The peasantry always "remains deeply subjugated". The difficulty is that Javeed fails to appreciate the true import of the concept of autonomy. When it is argued that the subaltern classes inhabit an autonomous domain, the implication is not that they are not dominated. On the contrary, it is precisely to conceptualise this domination as a relation of power that one must identify the autonomy of the subaltern classes. Domination must exist within a relation. The dominant groups, in their exercise of domination, do not consume and destroy the dominated classes, for then there would be no relation of power, and hence no domination. For domination to exist, the subaltern classes must necessarily inhabit a domain that is their own, which gives them their identity, where they exist as a distinct social form, where they can resist at the same time as they are dominated- It is only then that one can talk about domination as a relation, as a process, as a movement that emerges out of an opposition. To deny autonomy in this sense and simply to assert that the subaltern classes are "deeply subjugated" is to deny that they represent a distinct form of social existence; it is to merge their life into the life-history of the dominant classes. It would not do merely to add as a caveat that the oppressed too sometimes rise in "sporadic" revolts or that they are "not always manipulated" by dominant groups. The point is to conceptualise a whole aspect of human history as a history, i e, as a movement which flows from the opposition between two distinct social forces. Here to deny autonomy to the subaltern classes is to petrify this aspect of the historical process, to reduce it to an immobility, indeed to destroy its history. This precisely is what is done in elitist historiography, for there history moves either in terms of a unique bourgeois-feudal -opposition or, in countries like ours, a unique national-colonial contradiction. Nothing else matters. Javeed's misunderstanding of the concept of autonomy has another serious implication. He gives subaltern consciousness the peculiar construction of an "intermediate mental space" which lies between "the world of politics on the one hand and the economic processes of capitalist transformation on the other". How this space is structured he does not specify, except to say that it is confined within "the limitations of (an) archaic consciousness", that its outward expressions are "sporadic" and "spontaneous", and that in the process of capitalist transformation this space is restructured "in complex ways" by "a variety of ideologies and their variants" which turn a "pre-reflective critical consciousness" into something that is