The Semantic Role of Variability in the Development of Statistical Thought

Since the birth of modern sciences, the development of statistical thought has run along the evolution of the semantic concept of variability. The variability of the natural and social phenomena was the true challenge that Galilean science has faced substituting the order of scientific laws to the apparent disorder of facts. Those laws tried to combine two objectives: the explanation of phenomena in a causal context, and the forecasting of unknown events. In the twentieth century, the most revolutionary scientific theories have been very powerful as explanatory models, but weak as predictive models with reference to single events. All this because the new theories were first of all statistical ones, for example, the theory of evolution for natural selection, the genetics of population, or the quantum physics. Sciences learned to deal with statistic populations and collective properties. The intrinsic characteristics of this kind of laws were properties concerning a phenomenon as a whole, not its inessential micro components.