This article tries to assess the cognitive merits of various research programmes in the treatment of military issues. It contends that, on such topics more than on others, method influences conceptualization and theory construction. Having noted the infrequent showings of some programmes in the military field, it goes on to demonstrate, based on the literature produced over a century, that the blames laid at positivism's door are especially justified in it. The Weberian/Simmelian tradition has shown more promise. A brief synthetic formulation of its achievements suggests that it alone accords with Clausewitzian thought, and shows sufficient flexibility to account for recent changes which baffle other programmes, or to open avenues of collaboration with promising new approaches (notably, nonlinear dynamics).

One of the most notable political developments of the last thirty years has been increasing public and governmental demand for the quantification of social phenomena, yet sociologists generally have paid little attention to the spread of quantification or the significance of new regimes of measurement. Our article addresses this oversight by analyzing quantification – the production and communication of numbers – as a general sociological phenomenon. Drawing on scholarship across the social sciences in Europe and North America as well as humanistic inquiry, we articulate five sociological dimensions of quantification and call for an ethics of numbers.