DISCUSSION 61 of production but has to examine production in its tota}ity. The circularity of argument can be seen clearly in the discussion of the determination of the political and ideological instances in the last instance-the economic level, and on the base/superstructure metaphor (Social Scientist No. 141, p. 54-55). The reviewers' 'declaration' of the primacy of the economic instance is being taken to m^an that other levels were of "residual" importance. After the usual arguments about the mutual interaction and relative autonomy of th^ different instances of the mode of production Partha Chatterjee concedes that "if one insists", the economic level can be granted determination in the last instance. In the original review.^the work of Emmanuel Terray and Maurice Godelier had been used to show that the economic level was determinant in the last instance because it determined which of the instances of the mode of production was dominant. It is heartening to note that after his initial misgivings there is agreement between Partha Chatterjee and the reviewers on this point. As far as the categories base/superstructure are concerned it is rightly said that it is a "metaphor". The categories are used as a convenient shorthand for the economic instance on the one hand and the political-ideological instances on the other. Use of this metaphor does not mean reductionism in the least. Perhaps the word 'base' is associated with a'monolithic infrastructure of'the base'. The base or the economic infra structure as it is called, is itself a complex unity as outlined above,pf relations and forces of production. Anyway as Marx emphasised in the 'introduction' to The Grun-drisse^he point of departure for any analysis is social production which determines the other instances. After the declarations about the relative autonomy of the different levels in a mode ofproduction^he argument advanced takes an about turn and the reviewers are criticised for being hesitant to accept that "property relations are only the legal expression of the relations of production" and that "each can stand in the place of the others". According to Chatterjee, in the synchronous state of a mode of production "each instance corresponds to every others : They are in harmony" '(Social Scientist, No. 141 p. 5 6)—while in the transitional non-correspondence phase of a mode of production there are contradictions and antagonisms between the various levels. According to him the relative autonomy of the different levels of a mode of production exists only in the transitional phase of non-correspondence of the various levels. These formulations based on Etienne Balibar's contribution to Reading Capital are a distortion of his text. Balibar repeatedly reiterates that even in the correspondence phase,»the various instances or levels of a mode of pro-ductipn do not relate to each other as simple translation or transposition of one to another, but the correspondence between the levels is a complex set of interventions. The mode of'correspondence' between the different levels is the mode of articulation of these levels and not the expression or reflection of one in another. In fact,5eeing the relation between relations of production and legal forms as a relation ofinterchangeability is the worst form ofreduc-