ABSTRACT
The question I now must confront is how social scientific research can proceed in the absence of real possibilities of experimental control. How can the scientific enterprise even get under way without event regularities to account for? And how are competing theories to be assessed when the social sciences are generally denied the crucial test situation? Despite the lack of opportunities for controlled experimentation in the social sciences I remain optimistic about the social scientific prospects. In setting a context for developing my position I refer first, albeit extremely briefly, to relevant assessments of Bhaskar (1979) and Collier (1994). Each has questioned whether the potential for social scientific successes exists given the limited opportunities for meaningful experimentation in the social realm. And each has done so supporting a framework similar to that defended here. While Bhaskar sustains a degree of optimism Collier is rather pessimistic. A feature of social life that is fundamental to these opposed assessments is something which Bhaskar construes as a 'compensator' for the lack of experimental control in the social sciences. To the extent that Bhaskar is optimistic about the prospects for the social sciences this appears to turn on the existence of the suggested compensator. And it is through rejecting Bhaskar's arguments on this point that Collier concludes that any optimism concerning the possibilities of a successful social science are misplaced. I too find problems with Bhaskar's arguments, yet I do not share Collier's degree of scepticism about the consequences. I now run briefly though the various issues involved, and indicate the basis of my optimism. My reason for proceeding in this contrastive fashion is that I think it helps both to convey the central significance of the question posed at the beginning of this chapter and also to bring certain fundamental features of the argument into greater relief. These latter aspects include some upon which rather a lot ultimately turns but which appear not to be emphasised sufficiently by either Bhaskar or Collier.
