he
image of ourselves, not only because it is associated with us, but
especially because it is our natural complement. It becomes then a
permanent and integral part of self-consciousness to such an extent that
we cannot do without it and seek by every possible means to emphasize
and intensify it. We like the society of the one whose image haunts us,
because the presence of the object reinforces the actual perception and
gives us comfort. We suffer, on the contrary, from every circumstance
which, like separation and death, is likely to prevent the return or
diminish the vivacity of the idea which has become identified with our
idea of ourselves.

Short as this analysis is, it suffices to show that this complex is not
identical with that which rests on sentiments of sympathy which have
their source in mere likeness. Unquestionably there can be the sense of
solidarity between others and ourselves only so far as we conceive
others united with ourselves. When the union results from a perception
of likeness, it is a cohesion. The two representations become
consolidated because, being undistinguished totally or in part, they are
mingled and are no more than one, and are consolidated only in the
measure in which they are mingled. On the contrary, in the case of the
division of labor, each is outside the other, and they are united only
because they are distinct. It is not possible that sentiments should be
the same in the two cases, nor the social relations which are derived
from them the same.

We are then led to ask ourselves if the division of labor does not play
the same role in more extended groups; if, in the contemporaneous
societies where it has had a development with which we are familiar, it
does not function in such a way as to integrate the social body and to
assure its unity. It is quite legitimate to assume that the facts which
we have observed reproduce themselves there, but on a larger scale. The
great political societies, like smaller ones, we may assume maintain
themselves in equilibrium, thanks to the specialization of their tasks.
The division of labor is here, again, if not the only, at least the
principal, source of the social solidarity. Comte had already reached
this point of view. Of all the sociologists, so far as we know, he is
the first who has pointed out in the division of labor anything other
than a purely economic phenomenon. He has seen there "the most essential
condition of the social life," p