A Self-organizational Approach to the Gradual Learning Algorithm in Modelling the Evolution of Auditory Contrast

JAN-WILLEM VAN LEUSSEN & MAAIKE VONDENHOFF This report describes an attempt to remodel Boersma and Hamann (2007)'s Optimality Theoretic account of evolution of phonemic contrasts in an agent-based environment, inspired by agent simulations such as de Boer (2000). While we succeeded in transplanting B&H's model to the more complex world populated by agents, we were not able to replicate all their results in our model. Nevertheless we think our update of their model adequately explains some processes of language change, and introducing more features such as greater inter-agent differences would be worthwhile in modelling more complex cases of language change.