The argument put forward by this article is not that democratization does not benefit from the activity of a vibrant civil society, but rather that academic research should address this relationship in a critical way. This article maintains that one should take care to distinguish between
'civil society' as an ideal-type concept that embodies the qualities of separation, autonomy and civil association in its pure form, and the factual world of 'civil societies' composed of associations that embody these principles to varying degrees. At the same time, one should avoid a kind
of triumphalism about civil society as a necessary source of democratic energy with homogenous goals and principles; in a word, one should avoid a theory of civil society that privileges civil society (Fine 1997). A first problem seems to be mainly definitional: what is meant by civil society?
By reviewing the most relevant literature on democratization, the first part of this article discusses the main assumptions regarding the role of civil society as a democratizing power, namely its apolitical nature, its deep 'civil' stand and its relationship with the state. In the second
part, the article utilizes the case of South African civil society as a relevant example of how difficult and nuanced the relationship between civil society organizations and democratization can be, with special regard to the process of democratic deepening and social emancipation.
Theoria is an engaged, multidisciplinary and peer-reviewed journal of social and political theory. Its purpose is to address, through scholarly debate, the many challenges posed to intellectual life by the major social, political and economic forces that shape the contemporary world. Thus it is principally concerned with questions such as how modern systems of power, processes of globalization and capitalist economic organization bear on matters such as justice, democracy and truth.