eir political
opinions are non-rational. We can generally trust non-rational
inferences in ordinary life because they do not give rise to conscious
opinions until they have been strengthened by a large number of
undesigned coincidences. But conjurers and others who study our
non-rational mental processes can so play upon them as to make us form
absurd beliefs. The empirical art of politics consists largely in the
creation of opinion by the deliberate exploitation of subconscious
non-rational inference. The process of inference may go on beyond the
point desired by the politician who started it, and is as likely to take
place in the mind of a passive newspaper-reader as among the members of
the most excited crowd.

_(Chapter IV.--The Material of Political Reasoning, page 114)_

But men can and do reason, though reasoning is only one of their mental
processes. The rules for valid reasoning laid down by the Greeks were
intended primarily for use in politics, but in politics reasoning has in
fact proved to be more difficult and less successful than in the
physical sciences. The chief cause of this is to be found in the
character of its material. We have to select or create entities to
reason about, just as we select or create entities to stimulate our
impulses and non-rational inferences. In the physical sciences these
selected entities are of two types, either concrete things made exactly
alike, or abstracted qualities in respect of which things otherwise
unlike can be exactly compared. In politics, entities of the first type
cannot be created, and political philosophers have constantly sought for
some simple entity of the second type, some fact or quality, which may
serve as an exact 'standard' for political calculation. This search has
hitherto been unsuccessful, and the analogy of the biological sciences
suggests that politicians are most likely to acquire the power of valid
reasoning when they, like doctors, avoid the over-simplification of
their material, and aim at using in their reasoning as many facts as
possible about the human type, its individual variations, and its
environment. Biologists have shown that large numbers of facts as to
individual variations within any type can be remembered if they are
arranged as continuous curves rather than as uniform rules or arbitrary
exceptions. On the other hand, any attempt to arrange the facts of
environment with the same approach to continuity as is possible with the