dopt as a rule the paradox that an agreement proves more when it is
confined to a small number of circumstances. It is at such points of
coincidence between diverging statements that we are to look for
scientifically established historical facts.

(2) Before drawing any conclusions it remains to make sure whether the
_different_ observations of the same fact are entirely _independent_;
for it is possible that one may have influenced another to such a degree
that their agreement is inconclusive. We have to guard against the
following cases:--

(_a_) The different observations have been made by the same author, who
has recorded them either in the same or in different documents; special
reasons must then be had before it can be assumed that the author really
made the observation afresh, and did not content himself with merely
repeating a single observation.

(_b_) There were several observers, but they commissioned one of their
number to write a single document. We have to ascertain whether the
document merely gives the statements of the writer, or whether the other
observers checked his work.

(_c_) Several observers recorded their observations in different
documents, but under similar conditions. We must apply the list of
critical questions in order to ascertain whether they were not all
subject to the same influences, predisposing to falsehood or error;
whether, for example, they had a common interest, a common vanity, or
common prejudices.

The only observations which are certainly independent are those which
are contained in different documents, written by different authors, who
belonged to different groups, and worked under different conditions.
Cases of perfectly conclusive agreement are thus rare, except in
reference to modern periods.

The possibility of proving an historical fact depends on the number of
independent documents relating to it which have been preserved, and the
preservation of the documents is a matter of chance; this explains the
share which chance has in the formation of historical science.

The facts which it is possible to establish are chiefly those which
cover a large extent of space or time (sometimes called _general_
facts), customs, doctrines, institutions, great events; they were easier
to observe than the others, and are now easier to prove. Historical
method is not, however, essentially powerless to establish facts of
short duration and limited extent (those which are called _par