connected with the contemporaneous state of all the others. Religious
belief, philosophy, science, the fine arts, the industrial arts,
commerce, navigation, government, all are in close mutual dependence on
one another, insomuch that when any considerable change takes place in
one, we may know that a parallel change in all the others has preceded
or will follow it. The progress of society from one general state to
another is not an aggregate of partial changes, but the product of a
single impulse, acting through all the partial agencies, and can
therefore be most easily traced by studying them together. Could it even
be detected in them separately, its true nature could not be understood
except by examining them in the _ensemble_. In constructing, therefore,
a theory of society, all the different aspects of the social
organization must be taken into consideration at once.

Our space is not consistent with inquiring into all the limitations of
this doctrine. It requires many of which M. Comte's theory takes no
account. There is one, in particular, dependent on a scientific artifice
familiar to students of science, especially of the applications of
mathematics to the study of nature. When an effect depends on several
variable conditions, some of which change less, or more slowly, than
others, we are often able to determine, either by reasoning or by
experiment, what would be the law of variation of the effect if its
changes depended only on some of the conditions, the remainder being
supposed constant. The law so found will be sufficiently near the truth
for all times and places in which the latter set of conditions do not
vary greatly, and will be a basis to set out from when it becomes
necessary to allow for the variations of those conditions also. Most of
the conclusions of social science applicable to practical use are of
this description. M. Comte's system makes no room for them. We have seen
how he deals with the part of them which are the most scientific in