Page:Language and the Study of Language.djvu/117

III.] since, than as we pronounce them ourselves; and an orthoepical corruption or anomaly, like kyind for kind, dănce for dânce, neīther for nēither, is less frowned on by public opinion, and has a better chance for adoption into general use, than any, the most obvious, improvement of orthography.

The illustrations of phonetic change which we have been considering concern, as was claimed for them at the outset, only the most frequent and easily explainable phenomena of their kind, those which are found to prevail more or less in almost every known language. But every language has its own peculiar history of phonetic development, its special laws of mutation, its caprices and idiosyncrasies, which no amount of learning and acuteness could enable the phonologist to foretell, and of which the full explanation often baffles his art. His work is historical, not prescriptive. He has to trace out the changes which have actually taken place in the spoken structure of language, and to discover, so far as he is able, their ground, in the physical character and relations of the sounds concerned, in the positions and motions of the articulating organs by which those sounds are produced. He is thus enabled to point out, in the great majority of cases, how it is that a certain sound, in this or that situation, should be easily and naturally dropped, or converted into such and such, another sound. But with this, for the most part, he is obliged to content himself; his power to explain the motive of the change, why it is made in this word and not in that, why by this community and not by that other, is very limited. He cannot tell why sounds are found in the alphabet of one tongue which are unutterable by the speakers of another; why combinations which come without difficulty from the organs of one people are utterly eschewed by its neighbour and next of kin; why, for example, the Sanskrit will tolerate no two consonants at the end of a word, the Greek no consonant but n, s, or r, the Chinese none but a nasal, the Italian none at all: why the Polynesian will form no syllable which does not end with a vowel, or which begins with more than one consonant, while the English will bear as many as six or seven consonants about a single vowel (as in splints, strands, twelfths); why the accent in a Latin word has its place always determined by the quantity of the syllable before the last, and rests