Baseline probability analysis

Starting in 2012 we began development what we are calling baseline probability analysis (BPA)—the evaluation of historical, physical, biological, and social parameters having the prior probability of being relevant to many human situations, past and present.  

The essence of Bayesian logic lies in a simple proposition. Reflecting on what the word “plausible” means in the context of a statistical generalization, the Nobel Laureate Herbert Simon observed: “It means that the assumptions [being made] . . . are not inconsistent with our everyday general knowledge of these matters. At the moment they are introduced, they are already known (or strongly suspected) to be not far from the truth. The state of affairs they describe is not rare or surprising (given what we actually know about the world) . . .”.  

For example, before attempting to write about any given historical situation, problem, or event, it is sensible both to frame and also constrain the range of possible interpretations to those that are plausible given a reasonable set of likely prior probabilities.  We are proposing that such reasonable prior probabilities may be called baseline probabilities  

What Simon called limiting conditions may be similarly interpreted, although he may not have seen such conditions as operationally something to be determined prior to developing generalizations about the world, but we do.  

What we are suggesting, therefore, is that writing history must honor not only what are held to be “true facts” about the physical world, but also what are less openly acknowledged to be commonplace facts—baseline probabilities—about the biological, social, and historical properties of the world we humans inhabit and have created for ourselves. 

BPA can be used in conjunction with social network analysis (SNA) to identify, explore, and evaluate the robustness, or plausibility, of alternative historical claims and reconstructions.  

For example, systematic consideration of differing baseline parameters leads to significantly different interpretations of Lapita archaeology and the prehistory of the Oceanic (Austronesian) languages.

Regenstein Workshop on baseline probability analysis