most recognition is guided by such symbols.  One cannot make a new king, who may be a boy, in all respects like his predecessor, who may have been an old man.  But one can tattoo both of them with the same pattern.  It is even more easy and less painful to attach a symbol to a king which is not a part of the man himself, a royal staff for instance, which may be decorated and enlarged until it is useless as a staff, but unmistakable as a symbol.  The king is then recognised as king because he is the ‘staff-bearer’ ([Greek:  skeptouchos basileus]).  Such a staff is very like a name, and there may, perhaps, have been an early Mexican system of sign-writing in which a model of a staff stood for a king.

[11] Cf.  William James, Principles of Psychology, vol. ii. p. 392:—­’The whole story of our dealings with the lower wild animals is the history of our taking advantage of the ways in which they judge of everything by its mere label, as it were, so as to ensnare or kill them.’

At this point it is already difficult not to intellectualise the whole process.  Our own ‘common-sense’ and the systematised common-sense of the eighteenth-century philosophers would alike explain the fear of tribal man for a royal staff by saying that he was reminded thereby of the original social contract between ruler and ruled, or of the pleasure and pain which experience had shown to be derived from royal leadership and royal punishments, and that he therefore decided by a process of reasoning on seeing the staff to fear the king.

When the symbol by which our impulse is stimulated is actual language, it is still more difficult not to confuse acquired emotional association with the full process of logical inference.  Because one of the effects of those sounds and signs which we call language is to stimulate in us a process of deliberate logical thought we tend to ignore all their other effects.  Nothing is easier than to make a description of the logical use of language, the breaking up by abstraction of a bundle of sensations—­one’s memory, for instance, of a royal person; the selection of a single quality—­kingship, for instance—­shared by other such bundles of sensations, the giving to that quality the name king, and the use of the name to enable us to repeat the process of abstraction.  When we are consciously trying to reason correctly by the use of language all this does occur, just as it would occur if we had not evolved the use of voice-language at all, and were attempting to construct a valid logic of colours and models and pictures.  But any text-book of psychology will explain why it errs, both by excess and defect, if taken as a description of that which actually happens when language is used for the purpose of stimulating us to action.