We may start from the distinction between facts and our knowledge of them, for it is clear that whatever the character of the facts may be, our knowledge of them, at least, is progressive. The past is dead and gone. It is something over and done with, so that any change in it is forever impossible. We should then, if we would be precise, say, not that it is the past which grows and enlarges, but only our knowledge of it. We recover and conserve it in memory and imagination only, and as we recover it more and more completely and relate it more and more successfully, we know and understand it better. Plato is dead, and not one feature, circumstance, or action of his life can now be changed. He lives only in the memory of man; and because he lives there and stimulates the imagination, there is born a Plato of the imagination. There are thus two Platos, the one real and the other historical. The one lived and died long ago; the other still lives in human history. The real Plato has produced the historical Plato and affords a check upon historians in their representation of him. That representation may approach progressively nearer to what the real Plato was like, but it can never be the man who has passed away. History would be thus a branch of human knowledge, and grow with the growth of knowledge, while its objects remain unchanged. That is why history has constantly to be rewritten. Furthermore, in the rewriting, new types of history appear with new or altered emphases. The moral and religious type is supplemented by the political, and the political by the economic and social. For with the growth of knowledge the past looks different to us and we discover that what appeared once adequate has to be revised. We may admit, therefore, that history, whatever else it may be, is at any rate a kind of human knowledge. Like all knowledge it leads us to recognize that there is a distinction between knowledge itself and its objects, and that the progressive character of knowledge indicates an approximation to an adequate representation of the objects and not changes in their own character. This distinction in its application to history is evidently not a distinction between literature and its subject-matter. For the past, if we now take the past to be the proper subject-matter of written history, appears to have a twofold character. It is all that has happened precisely as it happened, and it is all that is remembered and known, precisely as it is remembered and known. There are, we may say, a real past and an historical past. The latter never is the former, but always a progressively more adequate representation of the former.