its origin is certain, not that its contents are free
from error. But authenticity inspires a degree of respect which disposes
us to accept the contents without discussion. To doubt the statements of
an authentic document would seem presumptuous, or at least we think
ourselves bound to wait for overwhelming proof before we impeach the
testimony of the author.

II. These natural instincts must be methodically resisted. A document
(still more a literary work) is not all of a piece; it is composed of a
great number of independent statements, any one of which may be
intentionally or unintentionally false, while the others are _bona fide_
and accurate, or conversely, since each statement is the outcome of a
mental operation which may have been incorrectly performed, while others
were performed correctly. It is not, therefore, enough to examine a
document as a whole; each of the statements in it must be examined
separately; _criticism_ is impossible without _analysis_.

Thus internal criticism conducts us to two general rules.

(1) A scientific truth is not established by _testimony_. In order to
affirm a proposition we must have special reasons for believing it
true. It may happen in certain cases that an author's statement is a
sufficient reason for belief; but we cannot know that beforehand. The
rule, then, will be to examine each separate statement in order to make
sure whether it is of a nature to constitute a sufficient reason for
belief.

(2) The criticism of a document is not to be performed _en bloc_. The
rule will be to _analyse_ the document into its elements, in order to
isolate the different statements of which it is composed and to examine
each of them separately. Sometimes a single sentence contains several
statements; they must be separated and criticised one by one. In a sale,
for example, we distinguish the date, the place, the vendor, the
purchaser, the object, the price, and each one of the conditions.

In practice, criticism and analysis are performed simultaneously, and,
except in the case of texts in a difficult language, may proceed _pari
passu_ with interpretative analysis and criticism. As soon as we
understand a phrase we analyse it and criticise each of its elements.

It thus appears that _logically_ criticism comprises an enormous number
of operations. In describing them, with all the details necessary for
the understanding of their mechanism and the reasons for their
employment, we are l