sight had not changed. Still less had the hieroglyphics gone out of use. The real and only cause of change was that the minds of the people had with the lapse of time become more accustomed to these characters, and that an abridged indication of them was sufficient to make them legible. It was therefore unnecessary to reproduce the detail of the figures, and all that was not essential was omitted. If the line of characters always presented an identical text, one of those unvarying formulas found in all languages, the abbreviations of the design would be still bolder. The hand and the eyes could easily run over these lines, the contents of which were known in advance.
The prime cause of phonetic changes is therefore mental. The word is a sort of vocal image impressed in the memory, the more or less complete reproduction of which is committed to our organs. The mind gradually familiarizes itself with this image, and no longer takes the same pains in reproducing it accurately, for it is sure of being understood. The will ceasing to watch over the organs, they follow their propensities. But if exactness becomes necessary, a slight effort of will is effective, the old consonants appear again, the contracted syllables resume their places, and we hear the word in its primary integrity.
While we have drawn our comparison from hieroglyphics, any movement directed by the will might have furnished a similar analogy. If we make the same gesture twenty times in succession, it will probably be less marked the twentieth time than the first.
Passing from one insensible change to another, it may happen that some sounds will quite disappear from the language, as has occurred, for example, with the liquid l in Zend, where it has been absorbed in r. If the organs in such cases seem incapable, it is not because they are different, but because they lack practice. If a Parisian youth is trained by an English governess, he runs the risk of having an English accent in speaking French. This does not prove that the conformation of his organs is peculiar, but that language has as much to do with making the organs as they with making language.
A third axiom is that the scale of sounds is never returned upon; that is, that when an articulation is once modified, it is never restored in its primitive purity. The habit of the Latin language is to contract its words; but domnus is at least as old as dominus, Hercles as Hercules, and valde as validus; and in inscriptions of the time of the empire, we find discipulina for disciplina, tempuli for templi, and liberitas for libertas. Change of s into r is one of the most general rules of Latin. But this change could not impose itself upon certain proper names, which fidelity or the taste for archaism
