Critical Realism and the Limits to Critical Social Science

Critical social scientists argue that explanations of social practices must be critical precisely in order to be explanatory, and that the necessity of critique gives social science a potentially emancipatory character. In practice, critiques of social phenomena are enormously contentious because it is difficult to establish agreement about what constitute problems, solutions or improvements, and whether the latter are feasible. The quest for emancipation therefore involves addressing normative questions and the feasibility of alternatives. Yet these considerations are unevenly represented in the practice of critical social science1 (hereafter CSS). They are most prominent in newer CSS’s such as feminism and green political theory, where discussions on topics such as the politics of difference or animal rights argue about the nature of problems and progressive alternatives. By contrast, until recently they were largely absent from the older Marxist-influenced CSS, reflecting the traditional Marxist antipathy to moral discourse and analysis of ‘blueprints’.2 The new CSS’s reflect the shift in politics away from the old certainties of class and distribution, where normative goals were largely taken for granted and the main concern was with attacking established power, to a politics in which debates about desirable alternatives are much more to the fore. There are now clear signs of a reawakening of interest among social scientists in ethics and political philosophy. For example, in development studies, critical social scientists who were previously sure of their value standpoints and realise they are problematic (Corbridge, 1993a and b; Toye, 1987). However, as they stand, the philosophical reconstructions of CSS are seriously out of kilter with these developments, and are implicitly closer to the older CSS in their neglect of the normative side (Fay, 1975; 1987; 1996; Bhaskar, 1986; 1991; Collier, 1994; Edgley, 1985). Critical realism in particular gives a complacent account of CSS in which ought follows straightforwardly from is,