approach scientific knowledge. Two presumptions, when they coincide,
make a double assurance. While memory, then, is the basis of all
historical knowledge, it is not called history until it enters a field
where it can be supported or corrected by evidence. This field is that
natural world which all experiences, in so far as they are rational,
envisage together. Assertions relating to events in that world can
corroborate or contradict one another--something that would be
impossible if each memory, like the plot of a novel, moved in a sphere
of its own. For memory to meet memory, the two must present objects
which are similar or continuous: then they can corroborate or correct
each other and help to fix the order of events as they really
happened--that is, as they happened independently of what either memory
may chance to represent. Thus even the most miraculous and direct
recovery of the past needs corroboration if it is to be systematically
credited; but to receive corroboration it must refer to some event in
nature, in that common world in space and time to which other memories
and perceptions may refer also. In becoming history, therefore, memory
becomes a portion of natural science. Its assertions are such that any
natural science may conceivably support or contradict them.

[Sidenote: Growth of legend.]

Nature and its transformations, however, form too serried and
complicated a system for our wayward minds to dominate if left to their
spontaneous workings. Whatever is remembered or conceived is at first
vaguely believed to have its place in the natural order, all myth and
fable being originally localised within the confines of the material
world and made to pass for a part of early history. The method by which
knowledge of the past is preserved is so subject to imaginative
influence that it cannot avail to exclude from history anything that the
imagination may supply. In the growth of legend a dramatic rhythm
becomes more and more marked. What falls in with this rhythm is
reproduced and accentuated whenever the train of memory is started anew.
The absence of such cadences would leave a sensible gap--a gap which the
momentum of ideation is quick to fill up with some appropriate image.
Whatever, on the other hand, cannot be incorporated into the dominant
round of fancies is consigned more and more to oblivion.

This consolidation of legend is not intentional. It is ingenuous and for
the most part inevitable. When w