representation are based on the so-called "representative
principle"--viz., that every section of the people is entitled to
separate representation in proportion to its numbers. The ideal varies
somewhat, but the usual conception, is that if each member represents a
different section or interest the assembly will represent all sections
or all interests. Now this is simply an attempt to return to what we
have described as the first stage of representation, but without the
fear of the monarchy to keep the sections together. For a deliberative
body or a king's council it might be suitable, but for an assembly
charged with the complete control of government in the interests of all
it is utterly impracticable. Each representative must represent all
interests; he must be elected on a definite policy as to what is best
for all the people. If he is sent in as the agent of one interest or one
section of the people, he ceases to be a representative and becomes a
delegate. All these schemes are therefore not proportional
representation at all, but proportional delegation.

We have shown that representation means the organization of public
opinion into two definite lines of policy, and that this is the only way
to prevent political anarchy. But the proportionalists (as they like to
call themselves) say that it means representing men and the opinions
they hold in proportion to their numbers. The fundamental error is that
they neglect the all-important factor of human nature. They look on
public opinion as something having an independent existence apart from
the questions about which it is expressed and from the means of
expressing it; and they fail to recognize that the character of public
opinion depends on the manner in which it is expressed and organized. It
is but a natural consequence that they also conceive the number of
sections of opinion awaiting representation as pre-existing and
independent of the electoral machinery.

In short, they reduce the whole problem to a nice little exercise in
mathematics, requiring only for its clear exposition some columns of
figures and a few coloured diagrams to represent the different shades of
public opinion. No better example of the dangers of _a priori_
speculation could be adduced than this chimerical idea of the
proportionalists that public opinion is something to be divided into
fractions like a mathematical quantity, unless it be, perhaps, the
conclusion that if you gather together