CHAPTER III.
NON-RATIONAL INFERENCE IN POLITICS The assumption--which is so closely interwoven with our habits of political and economic thought--that men always act on a reasoned opinion as to their interests, may be divided into two separate assumptions: first, that men always act on some kind of inference as to the best means of reaching a preconceived end, and secondly, that all inferences are of the same kind, and are produced by a uniform process of 'reasoning.' In the two preceding chapters I dealt with the first assumption, and attempted to show that it is important for a politician to realise that men do not always act on inferences as to means and ends.

I argued that men often act in politics under the immediate stimulus of affection and instinct, and that affection and interest may be directed towards political entities which are very different from those facts in the world around us which we can discover by deliberate observation and analysis.
In this chapter I propose to consider the second assumption, and to inquire how far it is true that men, when they do form inferences as to the result of their political actions, always form them by a process of reasoning.
In such an inquiry one meets the preliminary difficulty that it is very hard to arrive at a clear definition of reasoning.

Any one who watches the working of his own mind will find that it is by no means easy to trace these sharp distinctions between various mental states, which seem so obvious when they are set out in little books on psychology.

The mind of man is like a harp, all of whose strings throb together; so that emotion, impulse, inference, and the special kind of inference called reasoning, are often simultaneous and intermingled aspects of a single mental experience.
This is especially true in moments of action and excitement; but when we are sitting in passive contemplation we would often find it hard to say whether our successive states of consciousness are best described as emotions or inferences.

And when our thought clearly belongs to the type of inference it is often hard to say whether its steps are controlled by so definite a purpose of discovering truth that we are entitled to call it reasoning.
Even when we think with effort and with a definite purpose, we do not always draw inferences or form beliefs of any kind.