premise must be accounted for_. How are we able to
say that all men are mortal, and that lightning in the west is a sure
sign of rain? How was this general truth arrived at? There is only one
way, namely, through the observation of a large number of particular
instances, or through _induction_.

Induction is the method of proceeding from the particular to the
general. Many men are observed, and it is found that all who have been
observed have died under a certain age. It is true that not all men have
been observed to die, since many are now living, and many more will no
doubt come and live in the world whom _we_ cannot observe, since
mortality will have overtaken us before their advent. To this it may be
answered that the men now living have not yet lived up to the limit of
their time, and, besides, they have within them the causes working whose
inevitable effect has always been and always will be death; likewise
with the men yet unborn, they will possess the same organism as we,
whose very nature necessitates mortality. In the case of the
premonitions of rain, the generalization is not so safe, for there have
been exceptions. Lightning in the west at night is not always followed
by rain, nor can we find inherent causes as in the other case which
necessitates rain as an effect.

THE NECESSITY FOR BROAD INDUCTION.--Thus it is seen that our
generalizations, or major premises, are of all degrees of validity. In
the case of some, as the mortality of man, millions of cases have been
observed and no exceptions found, but on the contrary, causes discovered
whose operation renders the result inevitable. In others, as, for
instance, in the generalization once made, "All cloven-footed animals
chew their cud," not only had the examination of individual cases not
been carried so far as in the former case when the generalization was
made, but there were found no inherent causes residing in cloven-footed
animals which make it necessary for them to chew their cud. That is,
cloven feet and cud-chewing do not of necessity go together, and the
case of the pig disproves the generalization.

In practically no instance, however, is it possible for us to examine
every case upon which a generalization is based; after examining a
sufficient number of cases, and particularly if there are supporting
causes, we are warranted in making the "inductive leap," or in
proceeding at once to state our generalization as a working hypothesis.
Of course