On the status of linguistics as a cognitive science