, in the intervals of writing for the _Review_. In
these two months I completed the first draft of about a third, the most
difficult third, of the book. What I had before written, I estimate at
another third, so that one-third remained. What I wrote at this time
consisted of the remainder of the doctrine of Reasoning (the theory of
Trains of Reasoning, and Demonstrative Science), and the greater part of
the Book on Induction. When this was done, I had, as it seemed to me,
untied all the really hard knots, and the completion of the book had
become only a question of time. Having got thus far, I had to leave off
in order to write two articles for the next number of the _Review_. When
these were written, I returned to the subject, and now for the first
time fell in with Comte's _Cours de Philosophie Positive_, or rather
with the two volumes of it which were all that had at that time been
published. My theory of Induction was substantially completed before I
knew of Comte's book; and it is perhaps well that I came to it by a
different road from his, since the consequence has been that my treatise
contains, what his certainly does not, a reduction of the inductive
process to strict rules and to a scientific test, such as the syllogism
is for ratiocination. Comte is always precise and profound on the method
of investigation, but he does not even attempt any exact definition of
the conditions of proof: and his writings show that he never attained a
just conception of them. This, however, was specifically the problem,
which, in treating of Induction, I had proposed to myself. Nevertheless,
I gained much from Comte, with which to enrich my chapters in the
subsequent rewriting: and his book was of essential service to me in
some of the parts which still remained to be thought out. As his
subsequent volumes successively made their appearance, I read them with
avidity, but, when he reached the subject of Social Science, with
varying feelings. The fourth volume disappointed me: it contained those
of his opinions on social subjects with which I most disagree. But the
fifth, containing the connected view of history, rekindled all my
enthusiasm; which the sixth (or concluding) volume did not materially
abate. In a merely logical point of view, the only leading conception
for which I am indebted to him is that of the Inverse Deductive Method,
as the one chiefly applicable to the complicated subjects of History and
Statistics: a process di