SOCIAL SCIENTIST in other words does not refer to fixed globules into which society is partitioned, with class analysis simply taking the form of investigating how these fixed globules interact with one another. Class is not merely a concept used for obtaining a unique partitioning of society. It is a concept used for concrete, "practical" analysis. What is central, however, to the concept of class, no matter what the level of aggregation at which it is used, is that it is always defined with respect to the sphere of production. Even terms like "big", "small" and "middle" peasants are not just empirical classificatory terms; they are defined on the basis of criteria relating to the sphere of production. This emphasis on production derives from the materialist conception of history which is the basis of Marxism and whose starting point is the centrality of the man-nature dialectic. All production, through various layers of intermediation, represents the action of man upon nature. In the process of action however man learns the "secrets" of nature and uses them for his own purposes. This capacity to use the causality inherent in nature for his own teleology is basically what propels the development of productive forces. And since man's action on nature is not an individualist act but essentially a social act, an act conducted by men entering into definite social relationships with one another, the developing productive forces necessitate ever new forms of social relations which constitute the hallmarks of different modes of production. Now, the concept of class provides the possibility of the agency which enacts structural change. The Marxist perception of history in other words operates at two conceptually distinct but organically unified levels: at one level it is a story of structures getting transformed into other structures, of transitions between modes of production; at another level it is a story of agencies (viz. classes), coexisting with one another, struggling with one another, and effecting those structural changes which when viewed, as it were from "outside", appeared to be the abstract result of the conflict between the forces and the relations of production. The persistence of structures in other words is the result of a certain relation between classes; the transcendence of structures is also the result of a certain relation between classes. The concept of class therefore cannot be delineated in some arbitrary manner but has to be located within its specific structural context, which means in relation to the sphere of production. The foregoing has two implications which are worth noting. First, it is insufficient just to say that the history of recent societies is a