Abstrakt
The work analyses the underlying, implicit methods used by historical linguists. Five etymological dictionaries - Semito-Hamitic by Orel and Stolbova, Afroasiatic by Christopher Ehret, Nilo-Saharan dictionary by Christopher Ehret, Nilo-Saharan by Lionel Bender, Altaic Etymological Dictionary by Sergei Starostin et alii.
The implicit method of the authors was utilized on arbitrary corpora. We have shown that, if an arbitrary pair of languages, A-B, we have between 1 and 4 % matches - matches that are pure noise, simple coincidences - for a higher semantic tolerance and a bigger number of languages, the noise (number of spurious, coincidental matches) increases geometrically.
We show that most of the etyma discussed in these dictionaries are under the threshold of noise and should be considered as coincidences. We discuss, from a cognitive viewpoint, the biases that allow such errors to appear.
The pareidolia, Lake Wobegon effect, false consensus bias, Texas sharpshooter fallacy and representativeness heuristics and other cognitive biases in historical linguistics are illustrated and discussed. We show how the heuristics of the mind affects the reasoning - and historical linguistics reasoning.