Sounds are not assigned to ideas by pure chance. The choice is governed by the instinctive recognition of a certain logical relation between noises heard outwardly by man’s ear and ideas that his throat or tongue wishes to express. In the eighteenth century men were greatly struck by this truth. Unfortunately, it was caught in the net of etymological exaggeration so characteristic of the time; and its results were so absurd that they justly fell into disrepute. For a long time the best minds were warned off the land that had been so stupidly exploited by the early pioneers. They are now beginning to return to it again, and if they have learnt prudence and restraint in the bitter school of experience, they may arrive at valuable conclusions. Without pushing a theory, true in itself, into the realm of chimeras, we may allow that primitive speech knew how to use as far as possible the different impressions received by the ear, in order to form certain classes of words; in creating others it was guided by the feeling of a mysterious relation between certain abstract ideas and some particular noises. Thus, for example, the sound of _ē_ seems to suggest death and dissolution, that of _v_ or _w_, vagueness in the moral or physical realm, vows, wind, and the like; _s_ suggests starkness and standing fast, _m_ maternity, and so on.[117] Such a theory is sufficiently well founded for us to take it seriously, if kept within due limits. But it must be used with great circumspection, if we are not to find ourselves in the dark paths where even common sense is soon led astray. The last paragraph may show, however imperfectly, that material need is not the only element that produces a language, but that the best of man’s powers have helped in the task. Sounds were not applied arbitrarily to ideas and objects, and in this respect men followed a pre-established order, one side of which was manifested in themselves. Thus the primitive tongues, however crude and poor they may have been, contained all the elements from which their branches might at a later time be developed in a logical and necessary sequence. W. von Humboldt has observed, with his usual acuteness, that every language is independent of the will of those who speak it. It is closely bound up with their intellectual condition, and is beyond the reach of arbitrary caprice. It cannot be altered at will, as is curiously shown by the efforts that have been made to do so. The Bushmen have invented a system of changing their language, in order to prevent its being understood by the uninitiated. We find the same custom among certain tribes of the Caucasus. But all their efforts come to no more than the mere insertion of a subsidiary syllable at the beginning, middle, or end of words. Take away this parasitic element, and the language remains the same, changed neither in forms nor syntax.