DIALECTICS OF DECENTRALISATION 27 Senior Civil Society Adviser in the World Bank, for the fact that 'civil society5 and 'social capital5 are commonly used almost interchangeably in that, the most powerful of all international development agencies. The critique of the concept of social capital in itself starts to lay out an alternative approach which brings the state and political agency back in, and shows how 'public action5 - which I use as a label for the interplay of state and of non-state action in the public arena - may lead both to deepening democracy and to socially meaningful 'development5. What I think of as 'the dialectics of decentralisation5 may play a critical role in this process, as I believe the Kerala case demonstrates. 'DOWN WITH SOCIAL CAPITAL!1 'UP WITH POLITICS!' What a silly slogan this may seem. How can one possibly be 'against5 social capital, if it means, in Putnan^s words, 'networks, norms and trust that facilitate collective action5. And indeed my complaint is rather with the uses to which the term has been put than with the idea itself. In the hands of the two writers who more than any others have theorised social capital, Pierre Bourdieu and James Coleman, it is a useful but modest concept. Bourdieu speaks of social capital as meaning 'connections5 and he refers to more or less durable social relationships which are implicated in access to resources and hence, in the reproduction of class differentiation. According to this conception of it, social capital is certainly not a characteristic of 'society5 as a whole. No more is it in the work of James Coleman, even though his whole orientation, theoretically and politically, is very different from Bourdieu^. For him social capital refers to resources (of information and insurance) which arise from relationships between people, which are of value to them. There may be a public good aspect but social capital, in this conceptualisation, is not entirely fungible, and it is not at all evident that the social capital enjoyed by different groups of people can simply be added up into a 'total score5 for a society as a whole. In Putnan^s celebrated analysis of the performance of regional governments in Italy, however, it is suggested that whole societies may be differentiated in terms of the extent of their social capital. He aimed to demonstrate that the differences both in government performance and in levels of economic development between north central Italy and the Italian South can best be explained as being the outcomes of variations in 'civic engagement5, measured by political participation, newspaper readership and the density of horizontal,