A Simple Mathematical Law of Language Evolution
Last modified: October 24, 2007
With the passing of time, cultural traits and languages that are characteristic of human societies evolve. Old usages come to fall out of use as new rules emerge. While Proto-Germanic featured an elaborate system of conjugations to signify the past tense, its descendants Modern English, Dutch and German mainly use the dental suffix "-ed". The few irregular verbs that remain are fossils that testify to this evolution. Here we quantify the evolutionary dynamics that govern the emergence and spread of a dominant linguistic rule by studying the regularization of English irregular verbs. To this end we have generated a database of 177 English irregular verbs, tracking their inflectional changes over the past 1200 years. Only 145 of these were irregular by the year 1200, and 98 are still irregular today. We find that the regularization rate of irregular verbs is constant across the last 1200 years of evolution of the English language. A simple law governs regularization: the half-life of irregular verbs scales with the square root of their frequency of word use. A verb that is used 100 times less often is regularized 10 times faster. Using these observations, it is possible to make predictions about the future of the past tense in the English language.