Do such things as empty narratives and dead documents exist? In a certain sense, no, because external things do not exist outside the spirit; and we already know that chronicle, as empty narrative, exists in so far as the spirit produces it and holds it firmly with an act of will (and it may be opportune to observe once more that such an act carries always with it a new act of consciousness and of thought): with an act of will, which abstracts the sound from the thought, in which dwelt the certainty and concreteness of the sound. In the same way, these dead documents exist to the extent that they are the manifestations of a new life, as the lifeless corpse is really itself also a process of vital creation, although it appears to be one of decomposition and something dead in respect of a particular form of life. But in the same way as those empty sounds, which once contained the thought of a history, are eventually called _narratives_, in memory of the thought they contained, thus do those manifestations of a new life continue to be looked upon as remnants of the life that preceded them and is indeed extinguished. Now observe how, by means of this string of deductions, we have put ourselves into the position of being able to account for the partition of historical _sources_ into _narratives_ and _documents_, as we find it among some of our modern methodologists, or, as it is also formulated, into _traditions_ and _residues_ or _remains_ (_Überbleibsel_, _Überreste_). This partition is irrational from the empirical point of view, and may be of use as indicating the inopportunity of the introduction of a speculative thought into empiricism. It is so irrational that one immediately runs against the difficulty of not being able to distinguish what one wished to distinguish. An empty 'narrative' considered as a thing is tantamount to any other thing whatever which is called a 'document.' And, on the other hand, if we maintain the distinction we incur the further difficulty of having to base our historical construction upon two different orders of data (one foot on the bank and the other in the river)--that is to say, we shall have to recur to two parallel instances, one of which is perpetually referring us back to the other. And when we seek to determine the relation of the two kinds of sources with a view to avoiding the inconvenient parallelism, what happens is this: either the relation is stated to depend upon the superiority of the one over the other, and the distinction vanishes, because the superior form absorbs into itself and annuls the inferior form; or a third term is established, in which the two forms are supposed to become united with a distinction: but this is another way of declaring them to be inexistent in that abstractness. For this reason it does not seem to me to be without significance that the partition of accounts and documents should not have been adopted by the most empirical of the methodologists.