Science adds up for big softies

Making Social Science Matter

June 7, 2002

Harry Collins recalls his smart-ass days and considers the defeat of 'positivism'.

As a practical guide to newcomers to the social sciences, or as a corrective to those who think that more and better "objective" research will automatically turn social science into a clone of natural science, Bent Flyvbjerg's book is useful. He offers the old lessons with the freshness of someone discovering them for the first time. As a contribution to thought, however, the book took me back to the 1970s and arguments with writers such as Steven Cotgrove, author of the bestselling A-level text The Science of Society . Cotgrove provided the perfect foil for all us smart-ass phenomenologists, Winchians, Kuhnians, interpretivists, or whatever we styled ourselves, and we beat him up mercilessly. Flyvbjerg, unknowingly, has written another attack on Cotgrove.

But Cotgrove's book and the thinking behind it are no longer a force in most social-science circles. Indeed, the defeat of "positivism" has been so thorough that postmodernist carpet-baggers roam at will. Furthermore, we have learnt much more about natural science as a social enterprise and it looks much softer than before. Indeed, as we have looked closely at natural science over the past three decades it has begun to take on many of the features of the social sciences as described by us anti-Cotgrove smart-asses. For both these reasons, the argument for the old divide between natural and social science that Flyvbjerg reintroduces looks tired. It may even be time to reappraise what social scientists can learn from natural science.

The book starts with the Winchian-Wittgensteinian-phenomenological critique, which Anthony Giddens was later to refer to as the "double hermeneutic". Human actors live out their lives within a taken-for-granted reality that they construct collectively through their social life. My world contains mortgages, money, cricket and so on; the Azande's world contains witches, poison and I don't know what. To know what I might do next, you need to know about the framing of my world; to know what an Azande might do next, the first thing you must do is comprehend theirs. Thus, the social analyst has two sets of things to understand: the world created by the actors, and only then the world created by the analyst. The latter must be built on the former if it is to make any sense, which is the double hermeneutic. Since the analysts' conclusions are parasitical on the actors' framing of their worlds, and since social lives are bounded both in time and space, there can never be a universal predictive social science in the way there can be a universal and predictive natural science. All this could have been gathered from Peter Winch's Wittgenstein-inspired little book, The Idea of a Social Science , published in 1958.

Flyvbjerg then moves on to other writers, such as the Dreyfus brothers, Harold Garfinkel, Michel Foucault and Pierre Bourdieu. He wants to go further and show that even the kind of context-bound social science that it seems might be done within the double hermeneutic, is impossible. His main argument is drawn from the Dreyfus brothers' description of skill acquisition, which shows that many highly skilled persons cannot articulate the rules of their own performance. This, too, can be seen as an elaboration of the Wittgensteinian-Winchian analysis of what it is to follow a rule rather than a conceptual advance. For example, it is impossible fully to articulate the rules for continuing the arithmetical series 2, 4, 6, 8... We do not know that the next number is 10 from the unadorned rule "add 2", because that could equally give us 82, 28 and so on. To use the rule we have already to know how to use the rule for adding, and thus we find ourselves in a regress.

Spuriously, Flyvbjerg says that the fact that we cannot formulate the rules for the exercise of social skills means we cannot have a social science. But note that this would also rule out the science of arithmetic. Flyvbjerg's mistake is to think that one has to be able to spell everything out in order to have a science. But this is not and cannot be how science works. Fortunately for science, it does not need to grasp the essence of things; it need only relate the occurrence of one thing with another, or make some measurements or, in Paul Feyerabend's words, reach out for "anything that goes". This has been made clear to social scientists since the new analyses of science done in the 1970s and has been clear to thoughtful scientists since they discovered the mysteries of quantum theory. A context-bound social science can get by perfectly well, so long as it can measure and generalise the consequences of there being rule-following behaviour within contexts even if it cannot describe the essence of that rule-following.

More credibly, Flyvbjerg stresses the importance of values in social-science research and the central position of power. He lays out some rules for what he thinks of as a new kind of social-science research based on his own earlier well-known participatory case study of town planning in Aalborg, Denmark. He shows how he "discovered" the importance of power though his participation in this work; how he found that politicians tend to pick and choose between the findings of social scientists according to what fits their preferred policies. All this is already well known to social scientists but, again, it does no harm to have it rediscovered.

As a contribution to new thinking about social-science methodology, the best part is chapter six, Flyvbjerg's systematic defence of the case-study approach. That is nicely laid out. He also discussed the Aristotelian term phronesis , which is gaining salience. I am not sure what we are going to do with it, but one thing would be to see how it fits the new models of natural sciences.

Harry Collins is professor of sociology, Cardiff University.

Making Social Science Matter: Why Social Inquiry Fails and How it Can Succeed Again

Author - Bent Flyvbjerg
ISBN - 0 521 77268 0 and 77568 X
Publisher - Cambridge University Press
Price - £37.50 and £13.95
Pages - 204

Register to continue

Why register?

  • Registration is free and only takes a moment
  • Once registered, you can read 3 articles a month
  • Sign up for our newsletter
Please Login or Register to read this article.