This paper investigates the role of efficiency pressures in language evolution by examining universal attractors, defined as states around which language evolution revolves. Focusing on verbal person-number subject indexes in a worldwide sample of 383 languages, the study reveals an attractor characterized by specific lengths for each person-number combination, structured along Zipf’s predictions, and a strong preference for non-compositional, cumulative coding. The findings suggest that striving for less processing and articulatory effort are dominant efficiency pressures, although the latter is overridden by the need for constant information flow. Strive towards lower lexicon complexity and memory costs are weaker pressures for this grammatical category.

Ilja A. Seržant, George Moroz