Computational Simulation of Linguistic Change: An Agent-Based Model

Includes bibliographical references (pages 35-37).This study takes an individual-based approach on the role of social space and social network structure on the diffusion of linguistic innovation, a process that facilitates language change and variation. An agent-based model (ABM) that generates random graphs of interconnected collectives was devised to evaluate the effects of individual and collective variables on innovation adoption time. Significant findings include the phenomenon that individuals with relatively low degree ties (first through fourth) to the source of an innovation tend to adopt innovations much more quickly when they have a larger amount of weak ties. This supports a notion of a speaker-innovator who is a fringe member of a collective, characterized by a relatively large number of weak ties. This definition of the speaker-innovator is congruent with the "strength of weak ties" theory, which asserts that it is through weak ties that innovations travel beyond the innovator's local network to external networks. However, the model is not without weakness and several ways in which the model may be improved and extended are provide