24 
fully ascertained by a judicial examination of the evidence. 
Having now this set of agreeing instances, more or less 
numerous, which gives us, as it stands, only an induction per 
enumerationem simplicem, our task is, so to reason from it as 
to discriminate the propter hoc from the post hoc. The result 
of this task, when successfully performed, is to give us a 
“law of nature,” which is such because it is a law of true, 
efficient causation. It is to effect this we need the methods 
of logical induction. In stating them, the chief guide will be 
Mr. Mill, whose discussion in this point seems the most com- 
plete and just. 
As his excellent treatment has made the methods of induc- 
tion familiar to scientific men, little more will be needed for 
present purposes than the mention of them. 
1. The “ method of agreement”? is applied when in several 
observed cases a result, X, is preceded by different clusters of 
apparently immediate antecedents. In one instance, A, B, 
and C are observed to precede X; in another, A, D, and H 
precede X; in a third, A, F, and G precede X. On com- 
paring all the cases, we conclude that A was, all the time, the 
true, efficient cause of X, because it alone was present each 
time X arose. ‘The canon of the ‘‘ method of agreement,” 
then, is, Whichever of observed antecedents remains alone 
unchanged next before the effect is the true cause thereof. 
But this canon may yet fail to give us a demonstrated result 
(a), because a latent antecedent may lurk unobserved among 
A, B, C, D, E, F, G, not detected in either instance; (0) 
because one efficient may produce X at one time and another 
at a different time; and (c) two or more causes may have 
combined to produce X. 
2. “The method of difference” is applied to a set of 
instances when, if one of a given group of antecedents is pre- 
sent or is absent, a given sequent is correspondingly present 
or absent. A and B and C are followed by X and Y and Z. 
But when the antecedents are B and C (A being absent) only 
Y and Z follow, X being absent. A appears the cause of X, 
so far, that is, as we can know that the second group of ante- 
cedents, after which the one effect, X, failed, differed from 
the previous group only in the one circumstance, the absence 
of A, we know that A efficiently causes X. 
Yet the demonstration may not be exclusive, because A 
may be only one possible cause of X ; for often similar effects 
are the results of different causes, as heat results from chemi- 
cal reaction, or from electricity, or from percussion, or from 
compression, or from friction, or from vital energy. 
3. The method of “ agreements and differences ? combines 
