Numbers have become, over the last two centuries, a central feature of public discourse and a privileged means of founding knowledge and trust in many walks of life. GNP, the unemployment rate, the consumer price index, life expectancy, crime rates, estimates of risk, concentration ratios, IQ and a host of other statistical devices are now part of our informational environment – one might say of the very fabric of our daily existence. There are few problems we would now consider examining, discussing and tackling without making use of numerical knowledge and without, therefore, resorting to the rhetoric of numbers. This continuous expansion of statistical artefacts can be described as a momentous epistemic transformation that has to a large degree displaced earlier authoritative forms of persuasion based on local and singular types of knowledge, often couched in literary form, with information of a more general and standardized character, generally presented in a numerical guise. But this transformation has also been political and institutional, in the sense that it has been made possible because of the ongoing increase in information-gathering activities carried on by governments and their agents from the early nineteenth century on. The harmonization of weights and measures, the cadastral survey of the national territory, the advent of periodical censuses, the setting up of standardized procedures to record the basic vital events of each individual's birth and death, all partake to this multi-faceted enterprise, under the conjoined and yet distinct legitimacies of modern science and of the modern state.