Statistik

Statistics enters English from German Statistik, coined in 1749 by the German professor Gottfried Achenwall at the University of Göttingen to name a new academic discipline. Achenwall derived it from Modern Latin statisticum (collegium), meaning 'council of state,' itself from Italian statista ('statesman') and ultimately from Latin status ('condition, state, position'). Statistik, in Achenwall's formulation, was not the science of numbers — it was the systematic comparative description of states: their populations, resources, governments, military strength, economic output, and political characteristics. It was, in essence, political science conducted through the description of political units. The 'statistics' that Achenwall taught at Göttingen were closer to what we would call political geography or comparative politics than to what a modern statistician would recognize as their discipline.

The numerical side of statistics developed on a parallel track through a different tradition: the 'political arithmetic' of seventeenth-century English writers, particularly William Petty and John Graunt. Graunt's Natural and Political Observations on the Bills of Mortality (1662) analyzed the weekly death records of London to extract patterns — the ratio of male to female births, the seasonality of mortality, the distribution of causes of death. This was genuinely statistical reasoning in the modern sense: using numerical data to infer patterns about a population. Petty extended this approach to economic questions, estimating national income and population from indirect evidence. The word 'statistics' was not yet applied to this work; it belonged to Achenwall's descriptive political science. The numerical and the nominative traditions would not fully merge until the nineteenth century.

The merger happened through the Belgian astronomer and social scientist Adolphe Quetelet (1796–1874), who applied the techniques of astronomical error analysis to social data. Astronomers had long recognized that repeated measurements of the same physical quantity showed a characteristic distribution of errors around the true value — what Gauss had formalized as the normal distribution. Quetelet applied this insight to human measurement: heights, weights, and other physical characteristics of large populations distributed themselves according to the normal curve. He introduced the concept of 'l'homme moyen' — the average man — and argued that statistical regularities in human behavior were as real and as lawful as physical regularities. The state-measurer and the number-cruncher had found each other, and the combined discipline inherited the word 'statistics.'

The late nineteenth and early twentieth centuries saw the mathematical foundations of modern statistics constructed by Francis Galton, Karl Pearson, and Ronald Fisher. Galton developed regression and correlation; Pearson developed chi-squared tests and the Pearson correlation coefficient; Fisher developed the analysis of variance, the design of experiments, and the framework of significance testing and confidence intervals. These mathematical tools transformed statistics from the description of what is into the inference of what must be — from data-recording to hypothesis-testing. The word that began as a description of political states became the science of uncertainty itself, the formal method by which knowledge is extracted from incomplete data.