there is a tradition, and a true historical
record as well, the tradition is superfluous. When a tradition stands
alone, there is nothing to confirm it. What can we do then? To assume
the fact from the truth of the tradition, and the truth of the tradition
from the existence of the fact, is to argue in a circle. Two independent
traditions, however, may confirm each other. When this happens the case
is improved; but, even then, they may be but similar inferences from the
same premises.

If, then, I allow no inference which I feel myself justified in drawing
to be disturbed by any so-called tradition; and, if instead of seeing in
the accounts of our early writers a narrative transmitted by word of
mouth in lieu of a record registered in writing, I deal with such
apparent narratives as if they were the inferences of some later
chronicler, I must not be accused of undue presumption. The statements
will still be treated with respect, the more so, perhaps, because they
rest on induction rather than testimony; and, as a general rule, they
will be credited with the merit of being founded on just premises, even
where those premises do not appear. In other words, every writer will be
thought logical until there are reasons for suspecting the contrary. For
a true and genuine tradition, however, I have so long sought in vain,
that I despair of ever finding one. If found, it would be duly
appreciated. On the other hand, by treating their counterfeits as
inferences, we improve our position as investigators. A fact we must
take as it is told us, and take it without any opportunity of
correction--all or none; whereas, an inference can be scrutinized and
amended. In the one case we receive instructions from which we are
forbidden to deviate; in the other we act as judges, with a power to
pronounce decisions. Nor does it unfrequently happen that our position
in this respect is better than that of the original writer; since,
however, many may be the facts which he may have had for his opinion
beyond those which he has transmitted to posterity, there are others of
which he must have been ignorant, and with which _we_ are familiar.
Changing the expression, where there is anything like an equality of
_data_, the means of using them is in favour of the later inquirer as
against the earlier; in which case he understands antiquity better than
the ancients--presumptuous as the doctrine may be. With a _bona fide_
piece of testimony, however tradi