Evolutionary neural network model of universal grammar
Motohiro Saiki, Satoshi Matsuda · 2010
Acquisition and performance of languages or grammar are the typical intellectual activities of human beings, and various models of these processes using neural networks have been proposed. These activities, however, are considered not to be learned completely anew in each individual, but also to have been acquired over the long evolutionary history of human beings. The universal grammar is assumed to be a comprehensive knowledge of grammar that was acquired and hardwired in the brain during human evolution. By employing neuroevolution, we illustrate how the universal grammar might have evolved in the neural network using a genetic algorithm.