ABSTRACT
Since the late eighteenth century, social scientists and social reformers have generally considered that by relying on numbers they were following the model of the sciences of nature. This was at best only partly true. Administrative and mercantile uses of counts and measures go back at least as far as scientific ones. Book-keeping has as strong a claim as astronomy to be the prototype of quantitative reasoning, and the huge expansion of numerical information on populations and economies since about 1750 was largely directed by government officials with few if any scientific ambitions. For them, numbers were the quintessential form of neutral information, knowledge that could readily be communicated to great distances. Such information could also be easily summarized, combined, and rearranged, generally using operations no more demanding than the elementary operations of arithmetic. A tabular representation of births and deaths or commerce in each of the towns and provinces of a country, could be turned into a description of the whole simply by adding, supposing only that uniform categories had been used. Detailed verbal descriptions might be more informative, and of course in practice information would be reported in both verbal and quantitative form. But for the administration of large territories, it was convenient to rely on numbers whenever possible.
