Manipulation and Population Statistics in Nineteenth-Century France and England

Social Research, Summer, 2001 by Libby Schweber

In 1838, the Commission of the Academy of Sciences decided not to give the award to Demonferrand for his essay. While this decision could be read as evidence of the force of Moreau de Jonnes's epistemological or political arguments, an examination of the reasons for the decision and the follow-up suggests another reading. Instead of rejecting the work outright, the commission asked Demonferrand to resubmit it for the following year, along with a discussion of the procedures that he had followed in constructing his tables and the reasons for each step. It was on the basis of this expanded text that he was granted the award the following year.

In the report of 1838 the commission explained the reasons for the award. First it listed the different types of empirical finding--the new knowledge--the essay included. Then the commission insisted on the personal qualities this work had required, most notably the degree of effort involved. In the conclusion the commission wrote: "One can get a sense of the cares and pains [to which Demonferrand must have gone to get so many statistical calculations] when one thinks that Mr. Demonferrand based his calculations on an incomparably greater number of deaths than those of other savants who conducted similar research" (Comptes Rendus, 1838: 354) . Finally, the jury of the Prix Montyon divorced its evaluation of his practice from his calculations, which, it argued, were not relevant to the award of the prize. As the reporter explained: "It is not in the responsibilities of the Commission to verify calculations; it [thus] limited itself to recognizing the value of the methods which the author used in his choice and classification of materials. Such a vast work, which demanded so much research, and so many calculations, perfectly meets the intention of the founder of the prize, which was to advance statistics in France" (Comptes Rendus, 1838: 354).

This last comment concerning the responsibility of the commission to verify Demonferrand's calculations can be read in two different senses: either it referred to possible mistakes of addition or division or it referred to Demonferrand's mathematical calculations concerning the degree of error. In the first case, it situated the art of statistics in the laborious production of data, rather than in the work of calculation; in the latter it expressed the academy's more general distinction between mathematics and statistics.(9)

The commission's report thus points to a third epistemological model of statistical knowledge distinct from that of Moreau de Jonnes' or Demonferrand's. While Demonferrand presented his mortality tables as a contribution to a science that was approximate and abstract and whose authority rested on the use of mathematics to control for error, the academy awarded its prize for the quality of the data collected and for the new facts Demonferrand had disclosed. And while Moreau de Jonnes supported a descriptive model of statistics in which knowledge depended on a one-to-one correspondence with observation and on comprehensive national-level data that only the SGF could produce, the academy rewarded a second descriptive model whose authority rested on the personal qualities of the researcher. In terms of the closure of the debate, the institutional role of the academy as supreme arbitrator led to a reformulation of the debate as a controversy over types of evidence rather than types of sentence or epistemological models and a decision in favor of local-level data collection.