On the Biological Basis of Language Capacities