Computers that possess internal semantic networks can learn dialectically in the same way that humans do. Thus, while they do not intrinsically know how to add, they can learn.